Tribunal Criminal Tribunal for the Former Yugoslavia

Page 28229

 1                           Tuesday, 1 May 2012

 2                           [Open session]

 3                           [The accused entered court]

 4                           --- Upon commencing at 11.02 a.m.

 5             JUDGE KWON:  Good morning, everyone.

 6             Before we begin today's business, Mr. Tieger, as you know,

 7     Mr. Robinson has informed the Chamber that he would like his motion to

 8     exclude Sarajevo evidence to be resolved before the Prosecution case is

 9     closed.  While the Chamber does not necessarily agree that this matter

10     must be resolved before the Prosecution case is closed, it does see the

11     value in doing so if possible.  Thus, the Chamber would like to know how

12     this matter is progressing.  You told us last time that the meeting with

13     the Rule 70 provider was going to take place in the week of

14     17th of April.  Did that meeting take place?  If necessary, we may go

15     into private session.

16             MR. TIEGER:  Well, why don't we move briefly into private

17     session, Mr. President, just out of an abundance of caution.

18             JUDGE KWON:  Yes.  Could the Chamber move into private session

19     briefly.

20                           [Private session]

21   (redacted)

22   (redacted)

23   (redacted)

24   (redacted)

25   (redacted)

Page 28230

 1   (redacted)

 2   (redacted)

 3   (redacted)

 4   (redacted)

 5   (redacted)

 6   (redacted)

 7   (redacted)

 8   (redacted)

 9   (redacted)

10   (redacted)

11   (redacted)

12   (redacted)

13   (redacted)

14   (redacted)

15   (redacted)

16   (redacted)

17   (redacted)

18   (redacted)

19   (redacted)

20   (redacted)

21   (redacted)

22                           [Open session]

23             JUDGE KWON:  Yes, I shall repeat my question for you.

24             So my question relates to your Croatia binding order motion.  You

25     have now conducted two interviews, one with Mr. Tudjman and the other

Page 28231

 1     with Mr. Zagorec.  The Chamber would like to know what impact in your

 2     view this has on your binding order motion, Mr. Robinson.

 3             MR. ROBINSON:  Yes, Mr. President.  Having concluded those two

 4     interviews, we are in a position to withdraw the motion for binding order

 5     because we've been unable to establish that there are records in the

 6     possession of the Croatian government that have not been given over to

 7     us.

 8             JUDGE KWON:  So your oral withdrawal on this matter is final one?

 9             MR. ROBINSON:  Well, if we come up with subsequent information

10     that leads us to believe that they do have records, we might wish to make

11     a second binding order motion, but it's final in the sense that you can

12     take that off your books as far as we're concerned.

13             JUDGE KWON:  Thank you, Mr. Robinson.  The Chamber notes that

14     that motion has been withdrawn.

15             Having said that, shall we bring in the witness.

16                           [The witness takes the stand]

17             JUDGE KWON:  Good morning, Dr. Tabeau.

18             THE WITNESS:  Good morning.

19             JUDGE KWON:  Yes, Ms. West, please continue.

20             MS. WEST:  Good morning, Mr. President, Your Honours.

21                           WITNESS:  EWA TABEAU [Resumed]

22                           Examination by Ms. West: [Continued]

23        Q.   Good morning, Dr. Tabeau.

24        A.   Good morning.

25             MS. WEST:  Mr. Registrar, may we have P04995, please.

Page 28232

 1        Q.   Dr. Tabeau, when we left off last week we had just finished

 2     speaking about your Sarajevo reports.  And now we're going to begin to

 3     speak about the Srebrenica reports.  Now, the report you see on the

 4     screen before you, this is your report from 2009 on Srebrenica; correct?

 5        A.   Correct.

 6        Q.   And what was your task in writing this report?

 7        A.   This report was written to provide new evidence on the missing

 8     persons from the -- related to the fall of Srebrenica in July 1995 and to

 9     provide evidence about identifications of these individuals.  As

10     particular class of identifications that we were looking at were

11     DNA-based identifications produced by International Commission for

12     Missing Persons, ICMP, in Sarajevo.

13        Q.   All right.  And this is an older report.  This is not the most

14     recent one you've written; correct?

15        A.   Yes, this is a report from 2009, a report that was marking ten

16     years of research into victims of the fall of Srebrenica.  It is like, I

17     think, a ninth report in a series that we made.  There is a more recent

18     one which is however of a different type than what we see on the screen.

19        Q.   What was the main source used in writing this report?

20        A.   This report is about missing persons.  This is a very special

21     category of victims.  These victims cannot be reported in official death

22     notifications, as they are missing.  The source that was used for this

23     report is the ICRC list of missing persons, actually I call it a list but

24     it is a number of different lists compiled by ICRC.  ICRC is the

25     International Commission for Red Cross, an organisation mandated to

Page 28233

 1     collect information about persons going missing in all conflicts all over

 2     the world and to trace these individuals in all possible ways to make

 3     reports about the fate of these persons.

 4        Q.   And in the case of the events at Srebrenica, when was it that

 5     ICRC first started taking information?

 6        A.   ICRC started collecting the reports about missing of individuals

 7     directly after the fall of Srebrenica during the war in Bosnia and

 8     Herzegovina.  They were collecting information not only on Srebrenica but

 9     on the entire conflict, entire war in Bosnia and Herzegovina.  Srebrenica

10     was just this major episode, major incident that they were also covering.

11     They published their first edition of the list for Bosnia in 1996.

12        Q.   What other main source did you use in your report?

13        A.   To make clear first that the first edition of 1996 is just one of

14     many next editions published by ICRC.  So for this report, for 2009

15     report, we used the edition of 2005 and a separate list of ICRC missing

16     persons related to the fall of Srebrenica of October 2008.  The second

17     major source used for this report was the ICMP list of DNA

18     identifications of Srebrenica victims, a list that was issued in

19     November 2008.  And again, it is a list that has been updated

20     systematically by ICMP and recently we received yet another update, but

21     at that time we had the 2008 November update at our disposal.

22        Q.   Besides these two sources, did you consult other sources?

23        A.   Yes.  The methodology applied here in this report, for this

24     report, was our standard methodology.  That means sources like the 1991

25     census, voters' registers of 1997, 1998, and 2000 were used as a

Page 28234

 1     standard.  In addition to these two groups of sources we also used the

 2     official registration of internally displaced persons and refugees in

 3     Bosnia and Herzegovina as of 2000.  This is a source which we briefly

 4     called DDPR.  And we also used a number of smaller contextual sources.  I

 5     want to stress that in order to study the victims of Srebrenica we

 6     basically used on one hand ICRC records on the missing persons and on the

 7     other hand the ICMP records of DNA identifications.  These two were

 8     cross-referenced, compared, and through this we were able to conclude how

 9     many persons reported missing were as well identified by ICMP, meaning

10     that their bodies were exhumed from mass, entire, graves in the

11     Srebrenica area, their bodies were sampled, I mean remains were --

12     samples were taken from remains.  DNA profiles were made and compared

13     with the profiles of the surviving relatives.  So that's the core of our

14     work.  All other sources were used for certain important purposes.  Like,

15     for instance, the 1991 census was used as a major reference source to

16     validate information about missing persons.  Sources on displaced persons

17     and refugees were used to eliminate possible survivors.  Voters'

18     registers, 1997, 1998, 2000 with the same purpose, to find and eliminate

19     possible survivors.  And smaller contextual sources were used to compare

20     the results obtained from the major sources with those from smaller

21     sources.  What I mean contextual, we also used a source -- a smaller

22     source on displacement or on surviving population that became displaced

23     and et cetera.

24             THE ACCUSED:  Excuse me, just for the sake of public there was

25     no -- any translation in Serbian.

Page 28235

 1             JUDGE KWON:  From when did you not hear the translation,

 2     Mr. Karadzic?

 3             THE ACCUSED:  Her last few sentences.

 4             JUDGE KWON:  Just in case, could you repeat from voters'

 5     registers 1997.  From there could you repeat your answer?

 6             THE WITNESS:  Yes, yes.

 7             JUDGE KWON:  Just a second.

 8             Did you hear the translation of my intervention?

 9             THE ACCUSED:  Yes, now I hear but for few sentences, last

10     sentences, I heard only English.  And I suppose the public also didn't

11     hear.

12             JUDGE KWON:  Very well.  Could you repeat from there.

13             THE WITNESS:  So I am repeating my last answer.  In that answer I

14     summarised the main sources which we used to study victims of the fall of

15     Srebrenica were on one hand the ICRC list of missing persons and on the

16     other hand the ICMP DNA identifications of Srebrenica victims.  We

17     cross-referenced these two in order to draw conclusions about how many

18     individuals reported missing were later found in mass graves and other

19     graves in Srebrenica area and identified through the DNA analysis of

20     human remains and of surviving relatives.  Other sources that we used,

21     like for instance 1991 population census and voters' registers of 1997

22     and 1998 and 2000 were used as a standard -- as part of our standard

23     methodology.  The 1991 census to validate information about individuals

24     reported missing, voters' registers, in order to find possible survivors

25     and eliminate them from our list of victims.  Other sources, official

Page 28236

 1     records of internally displaced persons and refugees and similar were

 2     also used with the same purpose, to eliminate possible survivors.

 3             MS. WEST:

 4        Q.   Thank you.

 5             JUDGE KWON:  Thank you, doctor.

 6             MS. WEST:

 7        Q.   Doctor, now --

 8             JUDGE KWON:  Just a second.

 9             At the end of your previous answer you referred to smaller

10     contextual sources.  If you could explain it as well, just in case.

11             THE WITNESS:  Yes.  What I meant in this case we also used

12     additionally a set of databases provided from authorities in Bosnia and

13     Herzegovina also on displaced persons and refugees from particularly the

14     Srebrenica area.  We used them as additional source, not as the main

15     source, because the main source used was actually far broader and better

16     in terms of quality than this additional one.

17             JUDGE KWON:  Yes, Ms. West.

18             MS. WEST:  Thank you, Mr. President.

19        Q.   So, Dr. Tabeau, I want to now focus on the main sources which you

20     mentioned you cross-referenced for the purposes of this report, and

21     that's the ICMP list and the ICRC list.  What was the methodology you

22     used --

23             JUDGE KWON:  Just a second.

24             Mr. Karadzic, are you following the proceedings in the language?

25             THE ACCUSED: [Interpretation] Yes, Excellency, I am following,

Page 28237

 1     but I'm having problems related to my health condition due to this

 2     ventilation.  The air that is being blown onto my head is too strong and

 3     too cold and I cannot afford to suffer again.

 4             JUDGE KWON:  I would like the Registry to look into the matter

 5     again.  I thought that has been resolved, but -- just a second.

 6                           [Trial Chamber and Registrar confer]

 7             JUDGE KWON:  That will be resolved.  If you have -- are suffering

 8     from similar problem, please let us know.  In the meantime, we can

 9     continue.

10             Yes, Ms. West.

11             MS. WEST:

12        Q.   The methodology that you used, was this the same methodology you

13     used for your other reports?

14        A.   Yes, this is the same standard methodology, matching of sources

15     the individual record level which simply means that we trace individuals

16     between the lists we use in the analysis.

17        Q.   Now, we already spoke about that methodology last week and I'd

18     now like to look at the results.

19             MS. WEST:  If we can have 65 ter 23719, and this is the handout

20     that was given out last week, and specifically I'm looking at page 17.

21        Q.   Dr. Tabeau, do you have that?

22        A.   Yes.

23        Q.   It's on the very top we have table 1.  Can you tell us about this

24     table.

25        A.   This is a table summarising our update of the previous OTP list

Page 28238

 1     of Srebrenica missing.  The previous list is mentioned in this table as

 2     2005 OTP list.  On the top of this list there is a second category called

 3     2008 ICRC list, for which we only see 29 additional cases not reported in

 4     our 2005 list.  So what is shown here is how we critically looked at what

 5     we have done previously and expanded this result by using one more

 6     additional list of ICRC missing persons from Srebrenica.  This second

 7     list is mentioned as the 2008 list and relates exclusively to Srebrenica

 8     victims, exactly the same as 2005 list.  So the overall total of missing

 9     persons on the 2009 OTP list of Srebrenica missing is 7.692.  That is the

10     first type of information included in table 1.

11        Q.   And the second column of information that's labelled "number of

12     identified," does this refer to the ICMP list?

13        A.   Yes.  This refers to DNA identifications, positive

14     identifications of course, of persons reported at the same time as

15     missing individuals by ICRC.  So we cross-referenced it to, linked the

16     records representing the same individuals, and are able to say how many

17     of those reported missing at the same time have been identified by ICMP.

18     So the overall total of positive DNA identifications out of 7.692 was by

19     2009 5.061, which accounted for 65.8 per cent of the OTP list.

20             JUDGE KWON:  Ms. West.

21             MS. WEST:

22        Q.   Dr. Tabeau, if you move --

23             MS. WEST:  I'm sorry, Mr. President, do you have a question?

24             JUDGE KWON:  Yes, I was waiting for the French translation.

25             I wonder if Dr. Tabeau has explained to us the meaning of an OTP

Page 28239

 1     list.  I may have missed it.  How does it differ from ICRC report or

 2     something like that?  Yes, Ms. West.

 3             MS. WEST:

 4        Q.   Dr. Tabeau, can you talk about that, the difference between the

 5     OTP list and the ICRC list, if there's any difference at all.

 6        A.   The ICRC list, the core list that was used for the OTP list, is a

 7     broad list for the entire Bosnia and Herzegovina, and obviously the

 8     entire war period.  It is not what we need for the victims of the fall of

 9     Srebrenica.  So using this core list as the basis for our work, we

10     extracted relevant records using two criteria.  First of all, the place

11     criterion and second time criterion.  Places considered to be relevant to

12     the fall of Srebrenica comprised Srebrenica municipality itself and a

13     number of other municipalities in this area, in total ten municipalities,

14     which is all discussed in the annex, I think first annex, to 2009 OTP

15     report.  Time, of course, was also an important criterion.  The persons

16     who went missing in July 1995 and also later in the months between August

17     and until December 1995 were also considered relevant.  But as we will

18     see in few minutes there is a very clear time distribution of

19     disappearances that tells us that most of them occurred in July 1995.  So

20     it is all discussed in first annex to the 2009 report.  So the 2005 OTP

21     list or 2009 OTP list only means that we talk about a selection of ICRC

22     records relevant to the fall of Srebrenica and not about the entire

23     complete list for Bosnia and entire war period.

24        Q.   Dr. Tabeau, you had mentioned earlier that there was a

25     65 per cent overlap when you did your comparison, and we can see that on

Page 28240

 1     table 1; but can you now move down to table 12 and tell us how that table

 2     is connected.

 3        A.   This is a table that is much more complex than the first one.  It

 4     is much more complex first because it shows the result produced by myself

 5     and my team since the year 2000 up to 2009, but I suggest we focus right

 6     now on the last row in this table that is related to the results reported

 7     in the 2009 report.  And the second reason for the complexity of this

 8     table is that we go beyond the overlap of ICRC records and ICMP records,

 9     and in this table we show as well how many non-overlapping records were

10     reported by ICMP as positive DNA identifications of Srebrenica victims.

11     So if we go to the first column of this table in the last row we see the

12     familiar number, 7.692, it is what we just discussed, the same number

13     from table 1.  Next to it there is a column called "Srebrenica identified

14     (ICMP)."  This is a separate information summarising the progress ICMP

15     made in DNA identifications completely independently from the OTP

16     efforts, our efforts, to quantify the victims of Srebrenica.  So by 2009,

17     or rather, as based on November 2008, ICMP update in total they reported

18     5.555 positive identifications of Srebrenica victims.  It's a large

19     number and in the third column we again show how many of these 5.555

20     overlapped with the OTP missing persons.  So this number also familiar,

21     we just discussed it in table 1, is 5.061.  That's the overlap of two

22     independent sources reported -- reporting on Srebrenica victims.

23             JUDGE KWON:  Dr. Tabeau, yes, I have to admit this is rather

24     complex.  Still, we see the table 1 in this monitor.  We see the

25     "familiar" number 7.692 here, but there number of identified is 5.061,

Page 28241

 1     but in table 12 you said identified numbers are 5.555 and 5.061 is the

 2     number of overlap.  I have difficulty in following that.

 3             THE WITNESS:  ICMP works completely independently from the OTP,

 4     from ICRC, from anybody.  So once a while they issue a list of DNA

 5     identifications of Srebrenica victims.  This list issued at the end of

 6     2008 included 5.555 cases, persons.  This is a list that can be compared

 7     with all kinds of sources.  We compared this list with our list of

 8     missing persons from Srebrenica, from the fall of Srebrenica.  If you

 9     compare, you can see how many cases are identical, the same on both

10     lists, but you will also see that there are unique parts to the sources

11     compared.  So the unique part in ICMP records obviously was there as the

12     overlap identified is lower than 5.555.  It is not that we -- yeah, it is

13     just, you know, an exercise, you know, that we take independent sources

14     compiled by completely different groups and we did do it on purpose

15     because there is a lot of uncertainty in measuring or reporting victims

16     of war, and we want to make sure that all those who will be on our list

17     are also confirmed by other sources, by the ICMP in this case.  And

18     obviously next to 5.061 there are other categories in this table.  And if

19     we go, for instance, to the column called "new victims," identified

20     persons, new victims, this is one part of this unique part of ICMP, this

21     unique part of ICMP, that is not reported on our list of missing persons.

22     So they have extra 213 cases that we were unable to put on our list

23     because ICRC didn't have them.

24             JUDGE KWON:  Still difficult to follow.

25             THE WITNESS:  Your Honour, may I have --

Page 28242

 1             JUDGE KWON:  Just a second --

 2             THE WITNESS:  -- a graphical example.

 3             JUDGE KWON:  Just a second.

 4             THE WITNESS:  Yes.

 5             JUDGE KWON:  We should bear in mind the interval between the

 6     English and French translation.  We should put a pause.

 7             If you see the number 5.555, which is Srebrenica identified, so

 8     those parts that are not overlapping overlapped with the next number

 9     which is 5.061.  So those, about 500, are already identified.  Are they

10     not new victims, all of them?

11             THE WITNESS:  Not all of them, not all of them for a simple

12     reason that there is this matching process between ICRC records and ICMP

13     records.  For a group of records, 213, we are sure we cannot match them,

14     we cannot see any similar cases or the same cases on our list.  But there

15     is another category excluded overlap which actually relates to less

16     certain matches.  So we have made some matches but we are uncertain about

17     how well they represent our victims between these two lists.  So we

18     actually excluded them all together from the analysis because there is

19     uncertainty there and we leave this group for later.  Perhaps there will

20     be more results that will help us improve our matching in a later time.

21     But for now we know for sure that 213 are new.

22             JUDGE KWON:  Yes.  Yes.  I think I now understand the

23     mathematics.  So to put it -- I may put it like this.  Please correct me

24     if I'm wrong.  5.555 minus 5.061 equals 281 plus 213?

25             THE WITNESS:  Yes, that is correct.

Page 28243

 1             JUDGE KWON:  So if you could explain to us again when this

 2     overlap is to be excluded.

 3             THE WITNESS:  Overlap is excluded if we are not convinced that

 4     our matches are true matches, that they make it possible to say that two

 5     records, one from ICRC, one from ICMP, describe the same individual.

 6     There is a certain degree of similarity in these two compared records,

 7     but we cannot declare that this is highly reliable and highly likely that

 8     this is one and the same case.  So we just exclude them for now, but with

 9     time we expect there will be better information about these cases that

10     perhaps will make it possible for us to report them as true matches.

11     This kind of matching issue can be better understood if we think of ICMP

12     records with multiple names, first names, like, you know, if brothers

13     were killed or sons and father, sometimes they are reported jointly in

14     one record and this record has three different names.  In such cases ICMP

15     does not include at some stages date of birth of these victims, and for

16     us it is important to have the information about the date of birth, or at

17     least year of birth, in order to declare matches as good matches.

18             JUDGE KWON:  Does your report -- part of your report, main

19     report, deal with this matching exercise?

20             THE WITNESS:  Yes, yes, there is an annex which actually

21     discusses quite a degree of detail of our matching methodology in the

22     context of ICMP records.

23             JUDGE KWON:  Very well.

24             THE WITNESS:  If I may look for the number.

25             JUDGE KWON:  Yes, Ms. West, if you could be kind enough to show

Page 28244

 1     us an example how that matching exercise is being done in terms of this

 2     excluded, overlap, and new victims.

 3             MS. WEST:  Okay.  Just for the record, this is explained in

 4     annex 6.6 and that is page 102 of this particular report; that's the

 5     matching part --

 6             JUDGE KWON:  Third page from the last.

 7             MS. WEST:  And, Mr. President, when we look at the actual Excel

 8     spreadsheet we can give you an example of that.

 9             JUDGE KWON:  Yes.

10             MS. WEST:  And that will happen in one moment.  Thank you.

11        Q.   But, Dr. Tabeau, just to stay on this table so we can get to that

12     spreadsheet, I want to focus on the number 7905.  Please tell us what

13     that number represents.

14        A.   This number we can find in the last row in the column called

15     "Accepted victims (integrated)."  This is a third-last column.  It is an

16     important column because it integrates two numbers.  One is the number of

17     missing persons plus the number of new victims identified based on ICMP

18     identifications.  This is the number 213 reported in table 12 which we

19     just discussed.  So these two numbers, 7.692 plus 213 new victims gives

20     7.905, and this is the new overall total of Srebrenica victims on the

21     2009 OTP list.

22        Q.   When you gave that number back in your 2009 report, at the time

23     was ICMP giving you information that joined Srebrenica victims with Zepa

24     victims?

25        A.   Yes.

Page 28245

 1        Q.   So this number represents both Srebrenica and Zepa; is that

 2     correct?

 3        A.   Yes, it's correct.  It's mainly Srebrenica, though, because for

 4     Zepa the number at that time, as far as I remember, was 103 victims --

 5     well, generally Zepa is a totally different scale than Srebrenica.

 6        Q.   All right.  And when we talk about your most recent results we'll

 7     separate them as well.  But if you just look down to figure 1 you spoke

 8     earlier about time of disappearance.  Tell us about this chart.

 9        A.   This chart is showing the distribution of going missing by

10     calendar month in 1995.  The date of disappearance is reported by ICRC.

11     It is just one of the two major items, one is date of disappearance, one

12     is place of disappearance.  So from this chart, figure 1, we see that a

13     vast majority of victims disappeared in July 1995, as of course expected.

14        Q.   And the source of this information is ICRC; is that right?

15        A.   Yes, it's correct.

16        Q.   We can turn the page of the handout --

17             JUDGE KWON:  Just a second.

18             Can you go back to table 12.  What does the excluded potential

19     survivors number mean?  Does it have any impact on accepted victims

20     number?

21             THE WITNESS:  Your Honour, if I may, of course not.  The excluded

22     potential survivors are excluded and are not granted in the numbers

23     reported under accepted victims integrated.  They are shown here to

24     summarise our efforts related to our searching for potential survivors.

25     We found a number of such records, studied them, cross-reference in

Page 28246

 1     several sources like, you know, starting with the census, further voters

 2     register 1997, 1998, and 2000, records of displaced persons and refugees,

 3     and based on this analysis these 12 are excluded.  Once ICRC reports a

 4     very small cases alive.  So that means exclusions.  So one of these cases

 5     excluded is just this kind of ICRC alive person.

 6             JUDGE KWON:  So to understand it correctly, does this 12 -- I'm

 7     sorry, I correct.  Is this 12 included in Srebrenica missing number,

 8     i.e., 7.692?

 9             THE WITNESS:  It's not.  It's not included in any list of

10     victims.  This is just a separate category, separate category not

11     included in Srebrenica missing, not included in new victims, not included

12     in accepted victims, not included in ICMP identifications.  It is just

13     excluded.  It's not there.  Only exists as this separate number.  In the

14     list that is attached with the 2009 report these 12 victims are reported

15     in a separate annex as exclusions.

16             JUDGE KWON:  Then was this 12 included in the original ICRC list?

17             THE WITNESS:  At some point they were but not -- you know, ICRC

18     basically produces a list not only for Srebrenica but also for the -- for

19     other episodes.  Anyway, they were included but now we identified some

20     overlap with sources on survivors and they are excluded.

21             JUDGE KWON:  Thank you.  I was wondering if this number has

22     nothing to do with this Srebrenica missing, why this appears on this

23     table at all.

24             MS. WEST:  Thank you.

25        Q.   Dr. Tabeau, did you put this number here to show that you, OTP

Page 28247

 1     demographic unit, has done ongoing work with contextual sources to

 2     exclude any survivors?

 3        A.   Yes, this is to address these activities in my unit.

 4        Q.   So we were just talking about time of disappearance, figure 1.

 5     If we turn the page the next page in e-court we look at figure 3.  Does

 6     this regard place of disappearance?

 7        A.   Yes.  Figure 3 reports on the place of disappearance.  This

 8     figure shows a selection of places, the largest places reported as the

 9     place of disappearance.  It's rather a clear picture.  A few places, like

10     five places, including Potocari, forest, Kravica, Konjevic Polje, and

11     Kamenica, these five places alone are responsible for a very large

12     percentage of disappearances.  As a matter of fact, actually it is only

13     five municipalities where almost all these references can be located.

14     These places here in figure 3 are smaller than municipalities.  They are

15     concrete places, like Potocari such a place.

16             MS. WEST:  May we have 65 ter 11435 -- excuse me, it has a P

17     number already, P04996.  And if we can have e-court page 3 and 4.

18        Q.   Dr. Tabeau, attached to your report itself, did you also include

19     the list of names themselves?

20        A.   Yes, there is a long list of names attached with this report.

21     What I'm seeing on the screen actually is the first page of introduction

22     to this list.  This page explains the six components of the list.

23        Q.   All right.  And you had just mentioned when we were talking about

24     that excluded number of 12, excluded potential survivors, and we see on

25     the screen before us the parts, do those 12 people, are they listed in

Page 28248

 1     one of these parts?

 2        A.   Yes.  There is part 5 which is the list of exclusions, 12 cases

 3     reported in table 12.

 4        Q.   Now, if we looked at this in hard copy, this actual appendix is

 5     quite thick so we're not going to go through it, but can you tell us

 6     within these parts do you have an actual listing of the names that are on

 7     your missing list, missing persons list?

 8        A.   Yes.  This is what is listed here in parts 1, 2, 3, which are the

 9     major components of our list.  Part 1 is the records that enter our 2009

10     list from the previous, 2005 OTP list.  Part 2 are additional ICRC

11     records, the 29 that we discussed in the beginning of today.  Part 3 are

12     the new names of victims, the 213 names that we identified from the ICMP

13     records.  Part 4 are less likely matches, uncertain matches.  These are

14     the 218 which we don't count in our statistics as victims.  Part 5 are

15     exclusions that we don't count in our statistics on Srebrenica victims.

16     And part 6 are cancelled records of cases alive coming from ICRC as a

17     source.  So it is a very long list of almost 8.000 names that is attached

18     in this annex.  We list them using first name, father's name, surname,

19     date of birth, or just year of birth.  For ICRC records we are able to

20     tell exactly the date of disappearance, the place of disappearance.  And

21     for those cases that have been matched with the ICMP records, we are able

22     to insert additional information about the protocol number of this

23     victim, the protocol number meaning the reference number of this victim

24     identification report, and the name of site where the remains were

25     exhumed from.

Page 28249

 1        Q.   Thank you.  In your explanation you mentioned 218 which we don't

 2     count in our statistics as victims.  In fact, did you mean 281?

 3        A.   Yes, yes, absolutely.

 4        Q.   Thank you.

 5        A.   Thank you.

 6        Q.   I'd now like to move to the report -- the updated report for the

 7     Karadzic case.

 8             MS. WEST:  If we can have 65 ter 23710, please.

 9             Mr. President, for this report we will go through the

10     spreadsheet.

11        Q.   Now, you recently came to the OTP a few weeks ago, and when you

12     came what was your task?

13        A.   My task was to study the latest update of ICMP on Srebrenica

14     identified persons and cross-reference this update with the OTP records

15     of missing persons and assess the overlap between these two lists.

16        Q.   To assess that overlap, did you use the matching approach?

17        A.   Yes.  The methodology used in cross-referencing was exactly the

18     same as I have always used for these demographic reports.

19        Q.   We're now going to move to Sanction to show the spreadsheet and I

20     would ask that this not be broadcast.

21             MS. WEST:  Mr. President, I understand the decision from last

22     week in regard to the ICMP material.  That information is included here.

23     So for the time being I ask that this not be broadcast.

24        Q.   So, Dr. Tabeau, in front of you we see a spreadsheet, a

25     spreadsheet that the Trial Chamber is familiar with.  Can you tell us the

Page 28250

 1     origins of this?

 2        A.   The origin of this spreadsheet is the ICMP submission,

 3     January 2012 submission, on identifications of Srebrenica victims.  This

 4     is a selection of main cases only.  ICMP reports, actually both, the main

 5     cases and reassociations.  And for each main case -- for a main case

 6     there can be one, two, or even more reassociations, but in this

 7     particular spreadsheet if we go to the right a little bit then we will

 8     see that only main cases are included.

 9        Q.   Thank you.

10        A.   This one is from Srebrenica that we can see from the name of this

11     particular sheet at the bottom.

12        Q.   Now, on this spreadsheet the Trial Chamber sees two new columns

13     that were not on the original ICMP spreadsheet and that's column O and P.

14     Under O it says B-A-Z, BAZ (OTP 2009) integrated and then 2009 new & less

15     certain integrated.

16             Did you add this information to the spreadsheet?

17        A.   Yes, I included this extra items.  And they report on the results

18     of my matching exercise.  In the column O the so-called BAZ number is

19     reported or B-A-Z.  BAZ is the standard name for a record identification

20     number or case identification number in the ICRC lists.  Having this

21     number makes it possible to search through the ICRC lists, any list ever

22     published, and check information about this particular person, on the

23     circumstances and time of his or her disappearance.

24        Q.   So the BAZ number, is that just the ICRC identification number

25     just as protocol ID number is the ID number for ICMP?

Page 28251

 1        A.   Yes, this is this same kind of number.

 2        Q.   And under column O when we see a BAZ number, what does that

 3     indicate?

 4        A.   This means that a link has been established between the latest

 5     ICMP update of Srebrenica, the one from January 2012 and the ICRC records

 6     included on the latest OTP list.  The latest OTP list is the one of 2009.

 7        Q.   So any time the Trial Chamber sees that column filled in, does

 8     that mean that there's a match between ICRC and ICMP?

 9        A.   Yes, this is correct.

10        Q.   So now going back to the President's question in regard to how

11     you made the match, can you tell us, can you walk us through that process

12     by looking through this spreadsheet and I'll go back to the beginning of

13     it.

14        A.   So if we look at the information included in this spreadsheet we

15     have in column B the names of this particular victim included, that means

16     family name, father's name, and the first name of this victim.  In the

17     column C there is the date of birth, and this is -- these two together is

18     a good characterisation of this person which we can use to search for

19     this person in any other source, including the population census but most

20     importantly the ICRC records.  ICRC as well works with names, so also in

21     the ICRC if we search for this person there will be family name, father's

22     name, first name, date of birth, and additional information items related

23     to the disappearance of this person.

24             I want to draw Your Honours' attention to one thing.  We are

25     searching in a very limited environment in this particular case.  What I

Page 28252

 1     mean by the limited environment is that on one hand there is this

 2     detailed list of people selected from ICRC records of whom we know that

 3     they were reported as missing as Srebrenica victims and here is another

 4     limited environment.  This is the Srebrenica victims that is reported --

 5     that are reported by ICMP.  So having these two -- these two limited

 6     environments if we compared them, if there is a person born on the same

 7     or very close date because there are mistakes in that dates with the same

 8     names, although there might be differences in the spelling, it is a

 9     highly probable match that this is the same person represented in both

10     sources.

11        Q.   So if we go back to your column, every time -- under O every time

12     we see that BAZ number have you done a matching process and come up with

13     a link?

14        A.   Yes.

15        Q.   But now I also see in this case record number 3 under O there's a

16     blank.  Was there no link for that particular record?

17        A.   Yes, not in the ICRC component of the OTP list.

18        Q.   If we look at column P we see some information for that record.

19     It says 2009 ICMP new and there's a number.  Tell us about that

20     information.

21        A.   Yes.  So in case of no match with ICRC, there is still a

22     possibility that a victim was included in the 2009 OTP list.  In the new

23     records that were reported separately, in a separate annex, in a separate

24     list part 3 of our list.  So in order to distinguish between the matches

25     with ICRC and the new records on the 2009 list, in column P this

Page 28253

 1     information is inserted for the confirmed cases.

 2        Q.   Then do you also consider record number 3 to be a match?

 3        A.   Yes, the third line, but record number 2 --

 4        Q.   Sorry.

 5        A.   -- I would like to be specific.  Yes, this is the second person

 6     reported in this spreadsheet.  It is a match.  Yes, it is a match with

 7     the previously included new cases, the 213 cases.

 8        Q.   Now if everyone looks at the bottom of the screen we see separate

 9     tabs.  One's for Srebrenica, one's for Zepa, and one's for exclusions and

10     inconclusive.  Starting after your 2009 report, did ICMP start to deliver

11     the information on these victims separately, meaning victims for

12     Srebrenica, victims for Zepa?

13        A.   Yes, it was not that long ago that ICMP started to distinguish

14     between victims of Srebrenica and Zepa.  At the time of making the 2009

15     report this was not yet the case.  Only later, I think end of 2010 or so,

16     they started to -- presented this distinction.

17        Q.   All right.  So I've just clicked on the Zepa tab.  This regards

18     the list for Zepa victims; is that right?

19        A.   Yes, that's right.

20        Q.   Now I'm clicking on exclusions and inconclusive.  Please tell us

21     about this list.

22        A.   This is -- these are also new categories introduced not too long

23     ago but by ICMP.  They simply started reporting some exclusions or

24     inconclusive cases.  If we go to the right, a little bit to the right of

25     this spreadsheet -- so in the column ICMP comment we can read the

Page 28254

 1     justification of why this particular victim should be considered an

 2     exclusion.  This has nothing to do with DNA profiling and matching.  It

 3     has rather to do with the information that ICMP teams collected from the

 4     family members of these victims during the process of collecting blood

 5     samples.  There are a few cases of exclusions, I think in total nine

 6     cases and some more separately inconclusive cases, but it is always

 7     thought that the rationale for declaring a case an exclusion or an

 8     inconclusive case relates to this extra information obtained from the

 9     relatives of the victims.  Well, if we go again to the top to the column

10     R, to keep the column R, yes, then in the ICMP comment there is a case of

11     two brothers, one missing from 1992, another from 1995.  If we go to the

12     left then we will see that this is a record with two first names.  It is

13     two men reported jointly under this protocol number.  As it is often

14     thought that ICMP cannot distinguish on the basis of the DNA profile

15     between the two brothers, they don't know who is who, but they do know

16     from the relatives that one went missing in 1992, another in 1995.  But

17     what I can do in such situation, I can search for both brothers and I can

18     check which of the two is included as a Srebrenica missing person and I

19     can search for the other brother and I can make sure that the other

20     brother is not included in the ICRC missing persons.  This is what I did

21     in this case and obviously we have a record of the right brother on our

22     list and not of the wrong brother.  So if we create a link between the

23     right brother and this particular record we still counted this person

24     only one time, no more than just this one time.

25        Q.   So is it fair to suggest that these exclusions and inconclusive

Page 28255

 1     cases are ones that you specifically did some work on to make sure they

 2     were correctly included or not?

 3        A.   Yes, that is what it means.  I studied all these cases one by one

 4     and in the column match with 2009 OTP list and decision, these are the

 5     two columns, I summarised the results of my work.  All together I think,

 6     as far as I remember, I suggested five cases should be possibly

 7     reconsidered by the OTP or by myself and should be taken off from the OTP

 8     list.  And in total I think there were nine exclusions and 18

 9     inconclusive cases.

10        Q.   So now I'd like?

11             JUDGE KWON:  What is a PIP?

12             THE WITNESS:  PHR, I think or -- oh, PIP, yes, yes, yes.  PIP is

13     the short name for Podrinje Identification Project.  Podrinje

14     Identification Project is an organisation, as a matter of fact, operating

15     in Tuzla, in Bosnia-Herzegovina.  Actually it was established by ICMP and

16     funded long time by ICMP.  Now they are funded also by the government,

17     and are mandated to work on their identification of Srebrenica victims.

18             JUDGE KWON:  Thank you.

19             MS. WEST:

20        Q.   And now I'd like to go to the results.  If we can go back to the

21     handout to the last page, page 19.  65 ter 23719, page 19.  And here we

22     have two tables, and these are tables with which we're familiar.  It's

23     the same set up that we saw from your 2009 Srebrenica report.  And under

24     table 1, the second column, you see number of missing 2009 and for total

25     7692.  Is that the same number that we spoke about earlier?

Page 28256

 1        A.   Yes, it is the same table, only now it has been expanded.  It has

 2     two additional columns at the right side.  The second-last is called

 3     number of identified 2012 and the last column is per cent of the overlap

 4     between the 2012 update and the OTP list from 2009.

 5        Q.   All right.  If we can just do that in reverse.  We see on the

 6     bottom line 7692 and then 5061.  Those are from our earlier table;

 7     correct?

 8        A.   Correct.

 9        Q.   And the new number you just gave, 6241, is this the updated

10     number for the most recent ICMP list?

11        A.   Yes.  This number is the -- coming from 2012 ICMP update.  The

12     number 6.241 is the overlap of this latest update of ICMP with the number

13     7.692.  It used to be much lower.  It used to be 5.061 as of 2009 report.

14     Now it is 6.241, which gives us an overlap of 81.1 per cent.

15        Q.   And for those 6.241 records were lines that we would see on the

16     spreadsheet if we looked at it.  For each of those records would there be

17     a box filled in with a BAZ number or with a previous ICMP number?

18        A.   Yes, there would be such a box and in particular it would be the

19     BAZ box available for every of these 6.241.

20        Q.   If we can move below this to the next table, table 12, again this

21     is familiar.  But I want to go to the second-to-last column, new victims

22     identified 2012.  And we see a number there, 116.  What is that number?

23        A.   This is the victims that I was unable to confirm on any component

24     of the 2009 list of Srebrenica victims, and therefore they are listed

25     here as additional new cases from the 2012 update.

Page 28257

 1        Q.   And is this similar to what you did in 2009, where we see the

 2     column labelled new victims identified 2009 and you came up with 213?

 3        A.   Yes.

 4        Q.   So to get to the last column, 8.021 --

 5        A.   Yes.

 6        Q.   -- what numbers do you have to consider?

 7        A.   In order to obtain this number we have to take 7.905 previously

 8     accepted Srebrenica victims, once again 7.905, and on the top of it I am

 9     taking the 116 new cases from 2012.  And the result is 8.021 records

10     representing both Srebrenica and Zepa victims.

11        Q.   And do you know how many of the 2012 ICMP records regard Zepa?

12     Strike that.  Let me ask you another question.

13             If we take this number 8.021 you just said it includes Srebrenica

14     and Zepa.  If you subtract the Zepa cases, what's the number for just

15     Srebrenica?

16        A.   The number for Srebrenica, I think from this number we would have

17     to subtract 116 cases.

18        Q.   Okay --

19        A.   Coincidentally -- well, it's the same number as new cases, but it

20     is not the total new cases -- Zepa cases.  Well, once again, 8.021,

21     8.021, is the new overall number of Srebrenica victims and Zepa victims

22     jointly, obtained from the integration of the latest ICMP update from

23     2012 with the previously obtained number, which was 7.905.  These 8.021

24     cases covers both, Srebrenica, the fall of Srebrenica, and the fall of

25     Zepa.  Zepa is a small incident compared with Srebrenica.  In these 8.021

Page 28258

 1     cases there are only 116 cases that are related to Zepa.  It is

 2     consistent with the ICRC reports on the missing persons from Zepa.

 3     Already as of 1999 ICRC reported that the number of missing persons from

 4     Zepa was 118.  We have 116 out of 118 in our list.

 5             MS. WEST:  Mr. President, I'd like to tender the 2012 report,

 6     which is 65 ter 23710.  And I'd also like to tender the associated

 7     spreadsheet, and I would like that to be under seal for the time being,

 8     65 ter 23712.

 9             JUDGE KWON:  Before we do that, one further question about new

10     victims 2012, Dr. Tabeau.  You told us that they are the victims that you

11     were not able to confirm on any component of the 2009 list of Srebrenica

12     victims.  I don't think I understood that part.  Where did that come from

13     at all?

14             THE WITNESS:  These new cases come from the ICMP update of

15     January this year.  As previously, this update has been compared by

16     myself with the previous OTP list and the previous ICMP update.  So I

17     identified a number of cases that are not included previously at all, not

18     in the previous ICMP data, not in the ICRC used for the OTP list.  It is

19     just like new victims, which can be understood perhaps more easily if we

20     think about how is it possible that ICMP can report on new victims

21     different from ICRC reports.  Yes, it is possible.  It is that the

22     reports each of these organisations obtains goes -- is from the actual

23     families of the victims.  Some families contacted both ICRC and ICMP.  In

24     the case of ICMP they even donated blood for the identification.  Some of

25     them, however, a small number would go exclusively to ICRC and some of

Page 28259

 1     them would go exclusively to ICMP.  So if we think of how is it possible

 2     that, you know, there is still this new category, new records reported by

 3     ICMP, the answer is:  Yes, it is possible.  It is possible.  Which

 4     relates to the way these organisations operate and how in -- for months

 5     provide information to each of these organisations.

 6             JUDGE KWON:  So originally these 116 people were not included in

 7     the ICRC list?

 8             THE WITNESS:  Yes, this is what I'm saying.

 9             JUDGE KWON:  Thank you.

10             Any objections to the admission of these two reports,

11     Mr. Robinson?

12             MR. ROBINSON:  No, Mr. President.

13             JUDGE KWON:  Yes, we'll admit them both.  Putting the latter one,

14     the spreadsheet, under seal provisionally.  Shall we give the number.

15             THE REGISTRAR:  Yes, Your Honour.  65 ter number 23710 will be

16     Exhibit P5004.  65 ter 23712 will be Exhibit P5005 under seal.

17             JUDGE KWON:  Yes, Ms. West.

18             MS. WEST:  Mr. President, I'm changing subject matters at the

19     moment.  Is this the time for a break?

20             JUDGE KWON:  Yes.  We'll have a break for an hour and resume at

21     1.30.

22                           --- Luncheon recess taken at 12.26 p.m.

23                           --- On resuming at 1.34 p.m.

24             JUDGE KWON:  Yes, Ms. West, please continue.

25             MS. WEST:  Thank you, Mr. President.

Page 28260

 1        Q.   Dr. Tabeau, on Thursday you discussed a statistical formula, it

 2     was the capture recap -- is that working?

 3        A.   Yes, it's working now.  Thank you.

 4        Q.   So we were discussing the capture/recapture formula, which was a

 5     formula to estimate the overall size of a population when you couldn't

 6     actually count individual people.  You were asked by the Trial Chamber if

 7     you could come up with a more simplified form of the formula to make it

 8     more digestible.  Have you been able to do that?

 9        A.   Yes, I have.

10        Q.   Can you tell us about that.

11        A.   For this I need to use ELMO.  I prepared an example.

12        Q.   Thank you.

13             Go ahead.

14        A.   I am using an example of ducks which is because first

15     applications of this method were made among others for ducks.  Actually

16     in 1930 there was an interesting paper published for ducks.  So ducks is

17     an example of a wide population.  We don't know exactly how many are

18     there.  Sometimes we have the impression that there are very many because

19     they are extremely, extreme noisy.  And in order to know more about how

20     many are there outside in a park or in a natural area it is possible to

21     make an experiment which is designed to provide the unknown size of the

22     population of ducks.  I have on ELMO an example in which we see three

23     circles.  The biggest one is white inside and represents the entire

24     population of ducks living in a park.  This is what is unknown, how many,

25     and this is what we need to know.  In order to estimate this unknown

Page 28261

 1     number we take two samples from the population, two samples of ducks.

 2             The first one is the yellow one.  The yellow represents the first

 3     sample.  There are 200 ducks selected.  Every single of the 200 gets a

 4     tag on her leg so they are all marked.  All 200 have a tag that we can

 5     follow later and we let them go back to the wild.  So the proportion of

 6     the yellow circle to the big white circle is the first one that we need

 7     to keep in mind.  The sample size of the first sample, yellow circle to

 8     the big white circle, that is one proportion we have to think of.

 9             So sometime later we take another sample of ducks, this is the

10     green circle.  The green circle comprises 150 ducks.  When we catch them

11     we check whether they have this tag or not.  For a number of them we saw

12     that, yes, there were 30 of them with this tag.  The proportion of the

13     tag ducks that is represented in my drawing with the light green colour

14     to the entire sample -- second sample, which is 150 ducks, is the second

15     one that we need to remember.  The assumption of this method is that

16     these two proportions are equal, are the same.  That means the yellow

17     circle to the big white one is the same as the proportion of the 30

18     tagged ducks to 150 selected in the second sample.  It's just a simple

19     equality of two proportions that are used to draw the formula, which is

20     given below.  So N1 is the first sample, N2 is the second sample, N12 is

21     the overlap of the two samples, is the overlap of the two samples.  If we

22     make these two proportions N1 divided by N and we say it is equal to the

23     proportion calculated as N12 divided by N2, these two proportions give us

24     the formula at the bottom, which I will write down perhaps the two

25     proportions in order to make it more explicit.  The first proportion,

Page 28262

 1     yellow circle, N1 divided by the big white circle, divided by N, equals

 2     the light green area of the green circle divided by the size of the

 3     entire green circle.  After a simple change, transformation of this

 4     equality, we obtain this formula, from which we can calculate the unknown

 5     size of the population of ducks, which is in this particular case 1.000.

 6             There is also another rationale behind this method.  This other

 7     rationale originates from the probability theory.  In the blue box here

 8     I'm showing a very, very old formula for the probabilities of independent

 9     events, event A and event B.  So the probability of a joint occurrence of

10     two events, A and B, according to this formula, equals probability of the

11     first event, A times probability of the event B.  So the A and B events

12     are nothing different but these two samples, yellow and green circles on

13     my drawing.  And the probability of the joint occurrence of the event A

14     and B is an overlap of the two sources -- two samples.  It is the light

15     green area on the drawing.  And under the formula I just put the numbers

16     to express the probabilities.  And after again a transformation of the

17     numbers we end with exactly the same formula for the unknown N, as

18     previously discussed.

19             There is one important assumption to this method.  There are few

20     more but one is really very fundamental.  The two samples are

21     independent, which in case of the sources that were used in the Sarajevo

22     reports, one source being the Bakije funeral home and the second source

23     being the Sarajevo household survey, we can with confidence conclude that

24     these two sources were statistically independent because reporting a

25     victim in one source did not change the chance of the same victim being

Page 28263

 1     reported in the second source.

 2             JUDGE KWON:  Doctor, could you give me an example of sample in

 3     this case.  First N1 in this case, yellow ducks, could you give me an

 4     example of such sample.  This is not the whole ducks living in the park.

 5             THE WITNESS:  No, no.

 6             JUDGE KWON:  But could you give me some example to concretely

 7     understand that.

 8             THE WITNESS:  Yes.  So in order to create such a sample a person

 9     would have to go and catch the ducks one by one to end with 200 ducks in

10     the sample.  Having caught all these ducks, 200, we know for sure that

11     this isn't the complete population of ducks because there are many, many

12     more of course.  So that is how the sample was created.

13             JUDGE KWON:  Having caught these ducks from where?

14             THE WITNESS:  From a natural environment like a large park we can

15     think of, for instance, The Hague with all its parks as the environment

16     from which we will caught the ducks.

17             JUDGE KWON:  We are doing this exercise in order to know the

18     number of total ducks that are living in the park?

19             THE WITNESS:  Yes.

20             JUDGE KWON:  So a certain individual went to a park and caught

21     every duck that were in the park at that point of time?

22             THE WITNESS:  But if it would be all ducks --

23             JUDGE KWON:  All ducks, yes, yes.

24             THE WITNESS:  -- in that park it wouldn't be a sample.  It would

25     be perhaps a sample but representing a larger environment than just this

Page 28264

 1     park, right.  So the wide population we have to, of course, be specific

 2     about what is the wide population that we are approximating or

 3     estimating, you know, the size of.  In this particular example I am

 4     saying a park, what is this park, whether it is just one park in

 5     The Hague like, I don't know, Rosarium, for instance, park.  Rosarium

 6     park is just one of the parks in The Hague but there are more parks in

 7     The Hague that this area is kind of -- you can even work, you know, these

 8     green parts will take you from one park to the next.  So it is not just

 9     this park that this population of ducks is living in.  They will be

10     living in all parks in the neighbourhood.  This would be the natural

11     habitat for these ducks, and in order to know how many are there in this

12     natural habitat we have to estimate because we are unable to go and to

13     catch every single duck living there.

14             JUDGE KWON:  Very well.  If you could further expand on the

15     concrete example, green sample, that means there would be another park.

16             THE WITNESS:  [Microphone not activated]

17             JUDGE KWON:  Microphone, please.

18             THE WITNESS:  Okay.  Sorry, this was my mistake.

19             So the second sample is a second measurement.  So, as Your Honour

20     noted, at some point in time we go and take the first sample and then we

21     put them back and let them mix again with the remaining ducks.  So we

22     wait some time and we go again.  We can go to exactly same place.  So

23     from the same environment we pick another sample and it is justified to

24     expect that we will end with a different sample, largely different, not

25     entirely different though.  There will be this overlap, light green area

Page 28265

 1     which is marked here.  And the same, you know, it is a very good

 2     illustration of the relationships between sources on victims.  Sources on

 3     victims are very different from regular statistical sources like the

 4     population census.  Population census would represent the white circle

 5     here, right.  It would be all of the people.  But because of war the

 6     statistical system is most likely not operating the usual way.  We don't

 7     have all reports on what time deaths and especially about victims like

 8     missing persons of whom we don't have death certificates at the time they

 9     were gone missing.  So we have samples.  We have several samples, samples

10     that often largely overlap.  But there will be also this unique part in

11     every sample.  In this particular case the light green area is the core,

12     the overlap, and the dark green is the unique part to the second sample.

13     And the yellow part, non-overlapping part of the first sample is the

14     unique part of the first sample.  This is quite a normal situation in the

15     analysis of victims of conflict.  This is why it is very important to

16     analyse more sources than just one and check the overlap and check how

17     many are in the core, in the overlap, and how many victims are in the

18     unique parts.

19             JUDGE KWON:  Very well.  I take it then, Mr. Karadzic, will be

20     asking you many questions about the applicability of this principle to

21     the victims.

22             Mr. Karadzic, to understand you correct, although I note the

23     absence of your expert, Dr. Pasalic, are you challenging this principle

24     itself, capture and recapture principle in statistics field?

25             THE ACCUSED: [Interpretation] Absolutely.  I'm going to

Page 28266

 1     demonstrate why.

 2             JUDGE KWON:  Very well.

 3             Let's continue, Ms. West.

 4             MS. WEST:  Thank you, Mr. President.

 5             JUDGE BAIRD:  But, Doctor, before Ms. West proceeds, let me thank

 6     you very much indeed for simplifying the situation, making it more

 7     digestible and the ducks.  Thank you very much.

 8             THE WITNESS:  With pleasure.  Thank you very much.

 9             MS. WEST:  Thank you, Your Honour.

10        Q.   Dr. Tabeau, the other day there was the discussion about - you

11     were present for it - about certain items that had been disclosed and one

12     of the items that Mr. President asked about was your list, your flagged

13     list, of the 1991 census of those people who became displaced.  That

14     flagged list, is it your practice to disclose the results of the steps

15     you took along the way?  In other words, did you disclose that list to

16     the Prosecutors to hand over?

17        A.   Well, I think the practice in cases like this one is, first of

18     all, to provide all the sources that have been analysed in my reports and

19     provide reports discussing -- that are discussing the results.  In

20     addition to this, there is the practice of providing extensive annexes in

21     which details of the methodology are disclosed as well.  In addition,

22     there are references in terms of published books, articles, where the

23     same methodology is described.  I think I am sure that also in this case

24     this practice has been followed, and I am sure I remember myself certain

25     dead-lines from the time I worked in the OTP for disclosing this element

Page 28267

 1     I just mentioned.  And it was, for instance, for reports.  I had a

 2     dead-line in May 2009, early May 2009, to submit all the reports that are

 3     now presented in this case.  And at the same time I am sure that

 4     Mr. Karadzic received copies of everything.  It's been a long, long time

 5     that all these necessary materials were provided.

 6             MS. WEST:  Mr. President, in answer to your question the other

 7     day as to whether the flagged list was disclosed, the answer to that is

 8     no it was not.  However, both sources were disclosed.

 9             JUDGE KWON:  Thank you.

10             MS. WEST:  Thank you.

11             Mr. President, there are four addendums to the Galic report that

12     have not been admitted.  I've spoken to Mr. Robinson and the Registrar

13     has those numbers.  Mr. Robinson has no objection and I would ask -- I

14     would tender those four addendums as well.

15             JUDGE KWON:  Could you give us the --

16             MS. WEST:  Yes.

17             JUDGE KWON:  -- 65 ter numbers.

18             MS. WEST:  The first one is 65 ter 23695; second, 23696; the

19     next, 23697; and the last is 23698.

20             JUDGE KWON:  Thank you.

21             Mr. Robinson.

22             MR. ROBINSON:  That's correct, Mr. President, we have no

23     objection, but I hope you won't use that against us when Dr. Karadzic is

24     asking for a little bit more time because we are saving the Prosecution

25     some time in direct examination by doing that.

Page 28268

 1             JUDGE KWON:  That will all be admitted.  Let's give the numbers.

 2             THE REGISTRAR:  Very well they will be admitted as Exhibits P5006

 3     through P5009 respectively, Your Honours.

 4             JUDGE KWON:  Thank you.

 5             MS. WEST:  Lastly, Your Honour, I would tender the handout which

 6     is 23719.

 7             JUDGE KWON:  Yes, I think it may be convenient to follow the

 8     proceedings.

 9             Any objections, Mr. Robinson?

10             MR. ROBINSON:  No, Mr. President.

11             JUDGE KWON:  That will be admitted as Exhibit P5010.

12             MS. WEST:  And also, Mr. President, may the Prosecution upload

13     the diagram the doctor had presented in the last explanation and tender

14     that as well?

15             JUDGE KWON:  With Doctor's notation.

16             MS. WEST:  Please.

17             JUDGE KWON:  Yes, that will be admitted.

18             THE REGISTRAR:  As Exhibit P5011, Your Honours.

19             MS. WEST:  I have no more questions.

20             JUDGE KWON:  Thank you, Ms. West.

21             Yes, Mr. Karadzic.

22             THE ACCUSED: [Interpretation] Thank you, Excellency.  Good

23     afternoon, Excellencies.  Good afternoon to everyone.

24                           Cross-examination by Mr. Karadzic:

25        Q.   [Interpretation] Good afternoon, Dr. Tabeau.

Page 28269

 1        A.   Good afternoon.

 2        Q.   Once again I would like to express my gratitude to you for having

 3     met with the Defence team and I am sure it will be helpful in reaching a

 4     better understanding.  I am also thankful for your thorough approach in

 5     explaining the statistics to all the parties to these proceedings.  On my

 6     part I am going to do my best to put questions to you that can be

 7     answered with either a yes or no, and of course you are at liberty to

 8     choose which answer you're going to give me.

 9             THE ACCUSED: [Interpretation] Can we please now have the last

10     document put on the ELMO once again.  Thank you.

11             MR. KARADZIC: [Interpretation]

12        Q.   Dr. Tabeau, am I right if I say that this relates, these samples

13     and their interrelations, pertain to the park which is the largest

14     ellipsis of the three?

15        A.   Yes, this is true.

16        Q.   Thank you.  Is this park a contained or a closed system without

17     any entry or exit points?

18        A.   Well, for this particular population of ducks, they can move but

19     they would remain within the area studied as there wouldn't be birth and

20     migration out of the area or deaths.  The measurements were actually made

21     closely one after another.

22        Q.   But still within the confines of this particular park?

23        A.   Yes.

24        Q.   Thank you.  Would the result be affected in any way whatsoever if

25     the ducks were to move to another closed system, another park, which was

Page 28270

 1     not the subject of investigation and vice versa, if other ducks would

 2     migrate to this park?

 3        A.   If they moved to another environment, it of course will affect

 4     the estimation.

 5        Q.   Let me not even ask you about the love life of the ducks.  What

 6     if those ducks were to get drunk, to start killing each other, would that

 7     have any impact at all?

 8             JUDGE MORRISON:  Mr. Karadzic, it would have an impact on the

 9     ducks, but I mean a drunk duck or a dead duck is not alive and healthy

10     duck.  But it seems to me that what we should be concentrating on is the

11     mathematical principles and not the mating or drinking habits of ducks.

12             THE ACCUSED: [Interpretation] Excellency, let me remind you that

13     when talking about Freud you spoke about the love life of eels and the

14     principle is the same, is it not?  They're just nuances that have to be

15     altered.  Can this be applied to human population in view of the fact

16     that hazardous behaviour like drinking or murders or killing each other,

17     and so on and so forth, are things that are quite common among people

18     unfortunately.

19             THE WITNESS:  If I may answer this.  Now, first of all, Sarajevo,

20     the area studied in the two Sarajevo reports, where the estimation was

21     done also using capture/recapture, it was the closed population.  That is

22     very important observation that actually makes it possible to apply this

23     method.  It is a closed population in the sense that Sarajevo was blocked

24     for long time and it was not possible for the population to cross the

25     front lines to go out and to come back whenever they wanted.  So the

Page 28271

 1     population that we studied is a closed population, and this is the

 2     population that was questioned in order to provide reports about the

 3     victims in the Sarajevo households there.  Bakije operated within the

 4     same closed area and again collected reports on victims from dead

 5     population within the front lines, the population that was exposed to

 6     risk of being killed.

 7             Moreover, those who reported actually were the survivors and they

 8     were household heads who actually had a good overview of all what

 9     happened to these families during the conflict and during the time the

10     Sarajevo population was closed from the outside world by the front lines.

11     So from this point of view it is actually the assumptions of this method

12     are very well satisfied and I'm not worried about moving, migration,

13     these kind of things because these are not the sources that can give bias

14     to the capture/recapture estimates.

15             MR. KARADZIC: [Interpretation]

16        Q.   Thank you.  We're still talking about the principles of the

17     methodology, and what about this situation?  What if there were a tunnel

18     between this park and another park and what if there -- the tunnel were

19     used by thousands of ducks who were crossing from one park to the next?

20     Would that be something worth knowing?

21        A.   Well, there was a tunnel in Sarajevo, we all know that, and it

22     was used to provide aid to the people living in the city from the airport

23     where the aid was delivered.  Well, I'm not aware of huge, massive

24     migration through the tunnel out of Sarajevo, or, even more impossible,

25     into the Sarajevo that was under siege and attacked by shelling and

Page 28272

 1     snipers.  So I don't think it is a realistic scenario.

 2        Q.   I'm a bit confused, Dr. Tabeau.  Are you saying that food arrived

 3     in Sarajevo through the tunnel from the city airport?  Is that what you

 4     have just told us?

 5        A.   Well, I said there was a tunnel in Sarajevo connecting the part

 6     of the city under siege with the Sarajevo airport.  What was transported

 7     via the tunnel I am not aware in detail because I cannot be.  I wasn't

 8     there and I wasn't taking reports and I wasn't studying this part of the

 9     siege.  So the tunnel was there is a very well-known fact.  It has been

10     reported by media very frequently.  Even recently by BBC there is a

11     wonderful report about Sarajevo also about the tunnel there.

12        Q.   Thank you.  I wouldn't say that it was you who are to blame, but

13     it would be your hosts in Sarajevo who did not inform you about the fact

14     that the aid for Sarajevo passed through the Serbian territory, through

15     the streets, and that tunnel was under Sarajevo and that it connected

16     Hrasnica and Sarajevo and that you could easily enter Sarajevo or leave

17     Sarajevo and people indeed did that as much as they wanted to.  So there

18     was a connection between the two, but let's leave that aside and move on.

19             MS. WEST:  Mr. President, I would move to strike all of that.

20     It's a statement.  She already made a comment that she's not the person

21     to talk about what's coming in and out, and this is just a statement

22     without any question at all.

23             JUDGE KWON:  Thank you.

24             Let's move on, Mr. Karadzic.  [Microphone not activated]

25             MR. KARADZIC: [Interpretation]

Page 28273

 1        Q.   Then I have a question.  If it is correct what I'm saying, would

 2     you say that Sarajevo or the largest circle was as closed as you assumed?

 3        A.   Well, as I said, the population couldn't move freely across the

 4     front line and go to the rest of the country and back any time they

 5     wanted.  I am aware of it, but I didn't study it in my report.  So if you

 6     need evidence on the population movement in this period in this area,

 7     then I'm afraid it's not my reports that can provide you with this

 8     evidence.

 9        Q.   Thank you.  Just another subquestion.  On Thursday you mentioned

10     that you had visited the front lines in Sarajevo.  When was that?

11        A.   I think it was around the year 2001.

12        Q.   Thank you.  Did you see two positions along the front lines, the

13     Serbian position on one side and the Muslim position on the other side?

14     Did you see two parallel trenches running parallel to each other?

15        A.   At that time, you know, in 2001 around that time I couldn't see

16     trenches and parallel trenches.  It was long after the war ended.

17        Q.   However, if that was once a confrontation line, what were the

18     parties to that confrontation?  There must have been ones on one side and

19     the others on the other side; right?

20        A.   There are always parties to any confrontation, but I thought that

21     I'm here to speak about statistics and victims and not about parties to

22     conflict.

23        Q.   With all due respect, Dr. Tabeau, on Thursday you said that the

24     Serbs had been on hilltops from which they had a good view and they could

25     shoot from there.  However, if you had indeed visited the confrontation

Page 28274

 1     lines along the hilltops and if you could see that there was 10, 20, or

 2     50 metres between the trenches, then I would suggest that the Muslims

 3     were on the same hilltops; right?

 4        A.   Well, let's put it clear.  I am not a military expert and I would

 5     not comment on the trenches and positions of the armies.  It was -- it is

 6     not my area simply, you know.  It is -- I might have said that the Serbs

 7     were there on the hilltops.  That is a very well-known fact and I don't

 8     think I am making, you know, a discovery here, but that's all I can say.

 9     My goal, my objective and task, was to provide statistics on victims.

10     And moreover, I think when making my statistics I clearly distinguished

11     between civilian victims and military victims.  So it is not that I'm

12     saying that there were no military victims and that only exclusively

13     civilians were killed.  There were both, civilian and military victims.

14     It is all in my reports.

15        Q.   Thank you.  Should I be dealing with the matters that you

16     mentioned and that would fall under the category of explanations for your

17     statistical calculations in interpretations thereof and if that is taken

18     into account then I have to deal with that.  But if you were only dealing

19     with figures and somebody else should be dealing with implications then I

20     will abandon that topic.  But we will come to that.  Let me go on and --

21             JUDGE KWON:  Just a second.  Probably on Friday to my question

22     you answered this way -- let me put it this way.  I posed a question

23     about your conclusion about the deliberate targeting of the civilians,

24     and I asked you:

25             "Does the location of the army have any impact or effect on such

Page 28275

 1     result if the -- for example, if an army is located in the middle of

 2     civilian population.  Does it have any effect on that, Doctor?"

 3             And in the course of answering the question, you said:

 4             "I believe it does matter where the army is located ...  yes.

 5     For shelling there is a separate analysis in the report and I compared

 6     the victims of shelling for both civilians and soldiers, and the

 7     conclusion is not different for this particular category.  So these are

 8     two different patterns, I think.  As for the location of the army, the

 9     VRS was located, of course, on the hills with an excellent view on the

10     town down the valley, so I think it was a perfect location for shelling

11     of any victim."

12             Probably, Mr. Karadzic has in mind in posing those questions

13     whether such location has any impact upon your conclusion.  Any

14     observations, Doctor?

15             THE WITNESS:  I stand, of course, for everything I said last week

16     and what Your Honour just quoted.  And, well, what I tried to say is I am

17     not a military expert to comment on the position of the armies, parties

18     to this conflict.  That is one thing.  The fact that the VRS army was

19     located on the hills is a very well-known fact that I am sure also in

20     this case has been or will be or, I don't know, is discussed I believe by

21     military experts.  So in my view in my report what I am trying to do, I

22     didn't study the armies, their positions and their activities.  I studied

23     information on victims, on their causes of death, and on the patterns --

24     of the patterns of killings for both civilians and military victims.  I

25     compared the patterns and I draw conclusions as to whether these patterns

Page 28276

 1     were similar and proportional or not.  And based on the study of the

 2     patterns and causes of death my conclusion is that civilians most likely

 3     were targeted and there had to be a different force, a different

 4     mechanism, that was killing civilians and soldiers during the siege.

 5             JUDGE KWON:  Yes, Mr. Karadzic, please continue.

 6             THE ACCUSED: [Interpretation] Thank you.

 7             MR. KARADZIC: [Interpretation]

 8        Q.   We'll have to deal with your conclusions then, but let's complete

 9     our duck topic.  Would it be better if 200 were recaptured, for example,

10     and then to see what the overlapping was or does it make any difference

11     at all?  Can you just tell us briefly.

12        A.   Well, of course the numbers can change.  It is not -- that the

13     numbers I gave in the example are once and forever.  It does matter what

14     is the overlap and what are the unique parts, what are the sample sizes.

15        Q.   Thank you.  You provided an example here and you gave us the

16     months of March and April as temporal boundaries.  Is it significant at

17     all?  Is there a constant leading to the exact results when time

18     intervals are longer?  Would the time interval be significant at all when

19     we are talking about capture and recapture?

20        A.   Well, the March and April from the example were given to stress

21     that this is a separate moment at which these two samples were taken.  In

22     case of Sarajevo reports, one of the sources, Bakije funeral home, was

23     completed over a number of years starting in April 1992 and continuing in

24     the course of war and the other one was a one-time survey.  It was,

25     however, retrospective survey that was meant to collect information about

Page 28277

 1     all cases of killings and wounding since the beginning of the siege until

 2     the moment of the survey.  So I don't think that the March and April from

 3     my example do play any role in my actual work.  In my view these two

 4     sources that are used in the Sarajevo reports both describe the same

 5     population and there is nothing that could -- could be spoken up for one

 6     source being completely different and uncomparable with the first source.

 7        Q.   Would the result be different, for example, if we take the

 8     Sarajevo population and if the capture and recapture period were six

 9     months, a year, two years, or even three years?  How would that reflect

10     on the Sarajevo population?

11        A.   Well, I want to explain there is nothing like capture/recapture

12     period.  It is just the samples that I used for the capture/recapture

13     estimation.  What matters is what is the coverage of the samples, whether

14     they are comparable and can be cross-referenced, and my answer is yes

15     they are comparable, they can be cross-referenced, the overlap can be

16     identified, and the assumptions needed for this method are fulfilled.

17     This is why I used these samples in the capture/recapture estimation.

18             JUDGE KWON:  Just a second, Mr. Karadzic.

19             This will not be counted as your time, Mr. Karadzic.

20             Ms. West, with the assistance of Dr. Tabeau, could you show us

21     the page in her report that applies to capture/recapture to the Sarajevo

22     population.

23             MS. WEST:  [Microphone not activated]

24             JUDGE KWON:  Microphone.

25             MS. WEST:  Annex 6.6, I'll give you the page number.

Page 28278

 1             JUDGE KWON:  Could we upload that item.

 2             MS. WEST:  This would be in e-court -- this would be

 3     65 ter 12130.  E-court page is probably 58.

 4             THE WITNESS:  If we could move to the next page.  Table A6.1 is

 5     what we should be looking at.

 6             JUDGE KWON:  Yes.

 7             THE WITNESS:  This estimation is only made for the period from

 8     1st of April, 1992, to the 9 September 1992.  That's the Karadzic

 9     Sarajevo report.  In order to make this estimation records were extracted

10     from each source relevant only to this period.  We have Sarajevo

11     household survey 1.707 cases of Muslims both killed in natural deaths.

12     And then we haves move to Bakije and there is 1.576 cases of both again

13     killed and natural deaths of Muslims.  These two sources, samples, are

14     100 per cent comparable.  One was collected at mid-1994 backwards which

15     was retrospectively collecting all information about persons killed or

16     natural death during the siege.  And the same is with Bakije.  There is

17     nothing here like two periods or time interval or anything that would

18     diminish the correctness of these two samples and the possibility to

19     cross-reference them and compare it.  There is the overlap in the last

20     row of this table which tells 1.039 cases were reported involve samples.

21     So what Mr. Karadzic just told us about period of drawing the samples

22     does not apply to this situation at all.

23             JUDGE KWON:  1700 is a number of deaths that can be found in

24     household survey?

25             THE WITNESS:  Correct.

Page 28279

 1             JUDGE KWON:  And 1500 is number found in Bakije?

 2             THE WITNESS:  Correct.

 3             JUDGE KWON:  And there is an overlap of 1.000 numbers?

 4             THE WITNESS:  Yes.

 5             JUDGE KWON:  In that case, total number of deaths is 2.5 million?

 6             THE WITNESS:  No.

 7             JUDGE KWON:  No.  I'm sorry, 2500.

 8             THE WITNESS:  Yes, that is what we obtained too.

 9             JUDGE KWON:  Thank you.

10             Yes, back to you, Mr. Karadzic.

11             Thank you, Ms. West and Dr. Tabeau.

12             MR. KARADZIC: [Interpretation]

13        Q.   Could you please explain the first figure of 1.707.  Does it also

14     include Christian households, both Serbian and Croatian?

15        A.   Sarajevo household survey included all people who lived within

16     the front lines.  So obviously all ethnicities that were there are in the

17     survey records.

18        Q.   Thank you.  Would you say that Bakijes buried Christians?

19        A.   Of course not, it is a Muslim funeral home.  But for this reason

20     the whole capture/recapture started by taking Muslims from both sources,

21     so extracting relevant records from household survey Sarajevo as to end

22     exclusively with the Muslims and same with the Bakije funeral home.  I

23     want to mention ethnicity comes from the 1991 population census, from the

24     links with the census.

25        Q.   Thank you.  Do you know that in Sarajevo there is a funeral

Page 28280

 1     parlor which only deals with atheists, i.e., there's only atheists of all

 2     religions and ethnicities?

 3        A.   It might be so, but I didn't look for this and I didn't use this

 4     information.

 5        Q.   Let us say that the Bakije funeral parlor had a good record of

 6     all the bodies that they buried.  What about the Muslims that were not

 7     buried by the Bakije funeral home?  Is there comparable grounds or

 8     grounds to compare because there is a list of households, the second list

 9     is of the bodies that the Bakije funeral home buried, the third would be

10     a list of persons buried by other funeral homes?  How do you make your

11     calculations from that?

12        A.   We rather opted for using two sources of which we knew were very

13     well suited for these type of analysis.  Bakije is the largest funeral

14     home in Sarajevo with a long history and burying Muslims.  It is a very

15     good source for this kind of analysis.  So what we have done we have

16     completed the estimation for the Muslims and then extrapolated this

17     result over the other ethnic groups which is a very justified assumption

18     because we don't discuss here in the case of Sarajevo a case of ethnic

19     cleansing.  So there is a very good reason to assume that the same

20     extrapolation can be done for other ethnicities as what we produced for

21     Muslims from capture/recapture.

22        Q.   In your opinion is this formula applied to valid or invalid

23     variables?

24        A.   Well, of course it is in the case how we did it is applied to

25     valid variables, valid samples; otherwise, why would I be presenting it

Page 28281

 1     here?

 2        Q.   I'm sorry.  In my understanding a sample is one thing and a

 3     variable is another.  Is this applied to valid or invalid variables?

 4        A.   Well, then perhaps you tell me first what is a variable you are

 5     talking about and what is your definition of validity of this variable.

 6        Q.   We'll deal with that when the time comes.  Tell me, does this

 7     include any coefficient pertaining to age, average age of the population?

 8     Were death rates taken into account?  Could you apply any of that to this

 9     formula?

10        A.   I'm not sure what you are talking about, but I can tell that in

11     the capture/recapture estimation we didn't use any age distribution.

12        Q.   Thank you.  Am I supposed to understand this in the following

13     way, that this is not longitudinal, that it is not diachronic, and that

14     it has nothing to do with time?  Is that right?  What does that mean that

15     at one point in time there were 1700 and as far as Bakije is concerned,

16     1500.  Which period are you referring to?  What are the dynamics

17     involved?  What was it that you inferred on that basis?

18        A.   I mentioned the time-frame for this estimation earlier today.  It

19     is as in the title of the table A6.1.  It is the period from

20     1st of April, 1992, to 9 September 1992.  That is the time-frame for the

21     number, the estimated number produced.

22        Q.   Thank you.  Do you agree that perhaps this method could be used

23     by a bird institute or in the case of human beings an institute for

24     demographic research?  Do you believe that this kind of method would be

25     applicable in matters related to criminal law?

Page 28282

 1        A.   Actually I can tell you with certainty that capture/recapture and

 2     several other methods related to capture/recapture is one of the major

 3     methods for estimation of unknown numbers of victims of war.  It has been

 4     used in several other areas of study and there is a whole area of

 5     application called multiple system estimation where the number of samples

 6     goes beyond three and simple formulas like this one cannot be easily

 7     proposed.  So there is a log linear modeling analysis that was applied

 8     for this kind of study.  So I think capture/recapture is the way to go

 9     when it comes to unknown numbers of, among others, victims of war.  This

10     method has also been applied many times in epidemiology for estimation of

11     HIV patients, for instance, or for drug addicts that are also a category

12     that is very hard to measure and several other phenomena that cannot be

13     directly observed and measures.  So actually, what we have been doing

14     here is very much in line with a broader field of analysis related to

15     unknown population sizes.

16        Q.   Thank you.  Have you testified in any other court of law apart

17     from this one, and you testified here 17 times.  Did you testify anywhere

18     else?

19        A.   No, I didn't.

20        Q.   Thank you.  Except for epidemiological or demographic studies

21     that do not have criminal law applications, has this method been applied

22     anywhere else?  Has it been accepted in legal practice, as it were?

23        A.   I can't tell whether this method has been accepted in legal

24     practice related to epidemiology or other areas.  I think -- I wouldn't

25     think that in other international criminal tribunals a similar type of

Page 28283

 1     result were produced and tendered in evidence in other cases.  What we've

 2     been doing in this Tribunal is actually unique, I must say.  On the other

 3     hand, other tribunals, like for instance the Khmer Rouge trial

 4     in Phnom Penh, are in a much more difficult position because sources that

 5     are available for the estimation of war victims are not available to the

 6     same extent as it is the case in the former Yugoslavia.  From this point

 7     of view the war in Bosnia and Herzegovina is a very exceptional case.

 8     There is a lot of information, high quality often information that can be

 9     used to propose this kind of measures like counts of victims and

10     estimated more complete numbers.  It is a very unique situation, I think,

11     [indiscernible].

12        Q.   Thank you.  Now we're coming to the objective.  The objective of

13     this research of yours primarily dealt with Muslims, right, the

14     percentage of the Muslim population and displaced persons, refugees, and

15     so on; isn't that right?

16        A.   No, it is not right.  I think all ethnic groups were included in

17     my reports, especially in the municipality reports in which measures of

18     population displacement are presented, all ethnic groups are included.

19     The same is with Sarajevo, all ethnic groups are included and the same is

20     with Srebrenica.  Only in the case of Srebrenica it happens that almost

21     all victim were of Muslim ethnicity and were men.  So this is a result

22     but has nothing to do with our approach when it comes to ethnicity.  I

23     think in all the reports every ethnic group is represented.

24        Q.   Thank you.  We'll get to that.  So what was the objective?  Can

25     you tell us briefly what the objective was and also can you tell us

Page 28284

 1     whether you attained that objective.

 2        A.   Well, the objectives of course were different in all these

 3     reports.  In the municipalities report the objective was to measure the

 4     displacement of the population, the displacement related to the war in

 5     these territories.  Displacement in terms of both internal displacement

 6     and external displacement.  In Sarajevo reports the objective was to

 7     quantify the number of victims of the siege and analyse basic

 8     distributions of these victims.  The objective was as well to propose

 9     estimates that could be used to assess the degree of incompleteness of

10     the minimum numbers that were produced from our sources.  Finally, in the

11     Srebrenica reports the objectives are also different.  Srebrenica is a

12     very special case.  We study a particular group of victims, the missing

13     persons, of whom we have very good reasons to believe that they were

14     killed, they didn't die natural deaths as their bodies are often exhumed

15     from mass graves and other graves.  So the objective was to check, to

16     investigate, how many of them were there related to the fall of

17     Srebrenica and how much we know about the fate 15 years after the war

18     ended.  And obviously in the course of time we have learned a great deal

19     about the fate of these people, as today after 15 years we can say 81.1

20     per cent of all the missing persons have been identified.  We know the

21     DNA profile.  We know where the bodies were.  Families of many of them

22     were able to bury them in their own graves, family graves, so that that

23     were the objectives.

24        Q.   Thank you.  Since you've given such an extensive answer let us

25     try to complete one thing with regard to the things that you talked about

Page 28285

 1     today.  These 116 persons were added subsequently, like the 200-something

 2     who survived.  Were they at one point in time an integral part of the

 3     list of missing persons?

 4        A.   No.  The whole point is, as I discussed this several times today,

 5     sources on war victims overlap and largely so, but there are also unique

 6     parts of these sources that bring new information to the general picture

 7     of victimisation.  The new cases, 116, 213, are the unique records from

 8     the ICMP from DNA identification process of victims of Srebrenica.  For

 9     this reason they are reported as new and additional cases because we

10     couldn't confirm we couldn't find them in ICRC records.

11        Q.   And the persons who are now considered to be survivors, were they

12     ever part of the statistical mass of persons who were considered to be

13     missing?

14        A.   Yes.  At some point early in time they were, but the purpose of

15     our investigation of cross-referencing our list with sources on survivors

16     is to exclude such cases.  So in the course of time we were able to

17     confirm them as, for instance, records of refugees or displaced persons.

18     This is the reason that tells us something is wrong here and this is why

19     we have these 12 exclusions from our records.

20        Q.   What's the total number of survivors found in this entire mass of

21     almost 8.000 persons who are recorded as missing?

22        A.   Well, this is the 12 records that are listed in the table, so

23     that is the entire mass of possible survivors.  So the thing is that, you

24     know, if you work with records of missing persons, the hope that the

25     person at some point will be found alive is really very, very small.  So

Page 28286

 1     it is -- for missing persons what we see is as a general rule almost,

 2     these persons are found in the mass graves, other graves exhumed and

 3     identified.  It is for Srebrenica on the first place but also for other

 4     episodes of war.  So as I'm saying missing persons is a very, very

 5     special category of victims.

 6        Q.   Can we now say that the process of establishing the number of

 7     survivors has been completed or can new survivors still be discovered?

 8        A.   Well, I think that what could be done has been done, but every

 9     time when there are new victims, like the 116 or 213, for these new

10     records a search is always done and a comparison of these records with

11     sources on survivors, in order to see whether possibly for these new

12     cases there is an overlap with sources on survivors or not.  But because

13     these new records are coming from ICMP, that is, a source reporting on

14     DNA identifications of victims, again the chance that they would be found

15     on survivors lists is nil, simply not possible.  Sometimes it might

16     happen that the record of such a victim will be seen in the record of

17     refugees or displaced persons, but then it is very likely that a mistake

18     is in the survivor record.  It happens sometimes and it can be clarified

19     by questioning the family of such a person.

20        Q.   Thank you.  In your opinion, who reported missing persons for the

21     most part or in most cases or exclusively, if you will?  Actually, who

22     could have reported that a certain person was missing?

23        A.   ICRC, that is, the source that we used for the list of missing

24     from Srebrenica, have always been extremely selective when it comes to

25     the respondents.  The only accepted reports on a missing person from the

Page 28287

 1     family of this person, it is just the standard way they operate.  It is

 2     not possible for them to accept any report of going missing.  In order to

 3     produce a very reliable list of missing persons they have to follow very

 4     strict procedures about how to collect this information, and this is what

 5     they do.  So respondents must be the family of a victim.  Further

 6     information must be relatively detailed.  So it's impossible to drop just

 7     a name and that's it.  People have to be more specific about it.  ICRC

 8     uses a standardised questionnaire to collect answers from their

 9     respondents.  So there is a lot of attention paid by the ICRC when

10     collecting the data.

11        Q.   Thank you.  I'd kindly ask you to try to be as belief as

12     possible.  I'll try to put questions that can be answered by a mere yes

13     or no.

14             Was a mechanism established involving some obligation on the part

15     of a person who reported someone missing, did that person then have to

16     say if that person was found that the person was no longer missing?

17        A.   I am unaware of a mechanism that ICRC put in place.  It was, I

18     believe, up to the families to go back to the ICRC and tell them whether

19     or not this person was confirmed alive.

20        Q.   Thank you.  Is there a mechanism, or rather, is there an

21     obligation on the part of the family to do that every now and then, to

22     say that a person remains missing or to say that a person reappeared in

23     the meantime.  Or rather, to be more specific, once a person is reported

24     as missing for the first time, is there any established mechanism and

25     obligation on the part of the person who reported that to update the

Page 28288

 1     information provided?

 2        A.   I would say that probably there is no such obligation to the

 3     respondents.  However, ICRC co-operates closely with ICMP.  They exchange

 4     information.  And ICRC, as far as I know, is also updated by ICMP about

 5     their identifications.  By again comparing the lists, ICRC can categorise

 6     the missing persons into still missing persons, missing persons with a

 7     known place of death, missing persons dead, closed cases, and exclusions.

 8     This is actually how the ICRC provided us with their lists in 2005 and

 9     recently, more recently, in 2008.  So I would think there is an effort,

10     systematic effort at ICRC to update the information they have.  On the

11     web, for instance, only records of still missing persons are included,

12     not all the records, not the closed cases, but just still missing

13     persons.

14        Q.   Thank you.  I cannot give you an exact page reference now, but

15     I'm sure that you're going to recognise this assertion of yours.  When

16     you say that the reliability of sources is rather satisfactory but not

17     comparable to usual statistical sources, what did you mean by that?

18     Would it have been better had you had the usual statistical sources?

19        A.   The usual statistical sources that I used, by the way, in my work

20     as well, this is the 1991 population census that we used.  We also used

21     DEM-2 mortality databases that were compiled by professional

22     statisticians.  DEM-2 mortality databases were those used for the

23     Dragomir Milosevic report.  Well, what is different when a statistical

24     authority compiles a source, these are people who are trained to do this

25     work.  They know about how to collect information, how to develop

Page 28289

 1     questionnaires to be used in surveys, how to process this information.

 2     They know about cleaning the data.  They know about elimination about

 3     duplicates and about cases that are incomplete.  So if a statistical

 4     authority makes a source, obviously the quality of information is better.

 5     However, ICRC, not being a professional statistical authority, in my

 6     view, has done a lot of effort to produce reliable information.  And as

 7     far as I can compare, you know, ICRC versus other sources that I have

 8     been using, I can tell that ICRC is a reliable source of information.

 9        Q.   But do you agree that it would be a good thing if two phenomena

10     were to be compared from identical sources.  To be more specific, for

11     example, you compared the 1991 census and the voters register from 1997

12     and 1998.  Are they the same kind of sources?

13        A.   Of course not because voters register is just a source made for

14     different purposes.  However, in the case of Bosnia in 1997 and 1998

15     election, voters register is nothing else but a subset of the census

16     data.  Actually, the requirement for a voter to register to vote was that

17     the voter had to be enlisted in the -- on the 1991 census on the first

18     place.  So from this point of view for the purposes of my work here, I

19     mean the municipalities report, the voters register is a very good source

20     to use.

21        Q.   I kindly ask you to give me answers that are as brief as

22     possible.  Was the census compulsory?  Did a citizen have to be

23     registered in 1991?

24        A.   All persons living in Bosnia had to participate in the 1991

25     census.

Page 28290

 1        Q.   And what about voting, did a citizen have to vote?  Did a citizen

 2     have to apply to vote?

 3        A.   Well, no, it is a voluntarily process, the registration to vote.

 4     So the answer is no, not everybody had to vote, only those who wanted.

 5        Q.   Thank you.  Do you agree that the elections in 1996 were closer

 6     to the events of 1991 through 1995 than was the case in 1997 and 1998?

 7     Also, thereby they would have been a more accurate indicator?

 8        A.   They were closer in time, but not more accurate.  The 1996

 9     election is reported by OSCE in their reports as being characterised by

10     cases of fraud and misusing the voter's name to acquire majority in

11     certain municipalities.  Secondly, it is also reported that for this

12     particular election, 1996 election, that the list of eligible voters

13     contained mistakes.  It was sorted, not according moreover -- not

14     according to the JMBG but other characteristics.  And it was extremely

15     difficult for the voters to find their name on this list.  So in effect

16     of this procedure, very deficient procedure, many thousand of voters were

17     unable to vote and participate in the elections.  So there are very good

18     reasons for not using the 1996 voters register, and instead to use

19     1997/1998 which are known of being very good high quality with a very

20     high participation rate of 88 per cent.

21        Q.   But the OSCE was in charge of these elections as well, so are you

22     trying to say that in 1997 and 1998 there were more voters than in 1996?

23     That would be less favourable for those who are championing genocide.

24        A.   Well, if we -- so the 1996 election register could not be used

25     for very good reasons.  So we had to use 1997/1998.  Whether the numbers

Page 28291

 1     produced would be better from the 1996 voters' register, I don't think

 2     so.  If you think about the process of returns then you can easily

 3     imagine that between 1996 and 1997 some people must have returned to

 4     their homes.  So this only means that they are not listed in our

 5     statistics as displaced persons.  So if you think how conservative or not

 6     conservative our statistics based on the 1997 voters' register are, I am

 7     telling they are more conservative because of the returns process.

 8        Q.   But why did you decide on that?  Why didn't you carry out an

 9     analysis including 1996 and then leave it up to the Chamber to decide?

10     Who made a decision to exclude the 1996 election data?

11        A.   Well, I think the decision was made in the demographic unit here

12     in the Tribunal and we in co-operation with the OSCE who actually

13     provided us with the voters' records.  And I think my role is not to

14     propose and make analysis that by definition are wrong and biassed.

15     Instead, my objective is to make and present analyses that are reliable

16     and meaningful.  So using the 1996 voters' register is not consistent

17     with this objective.

18             JUDGE KWON:  However, as a matter of fact, were there more voters

19     in 1996 election than in 1997 and 1998 election?

20             THE WITNESS:  I don't think there were more because many were

21     unable to register.  And moreover, there were cases, and frequently so,

22     of fraud, that voters registered for certain municipalities and they

23     shouldn't.  There were like duplicate registrations which inflated the

24     numbers in the register.  So these are serious biases and, well, it's not

25     a good source to be used in this kind of analysis.

Page 28292

 1             JUDGE KWON:  In answering the question because the Chamber does

 2     not know which numbers so it's very difficult to follow.  So in terms of

 3     result 1996 election has more voters?

 4             THE WITNESS:  I don't know exactly.  I can check the number to

 5     tell whether more or not.  Disregarding the number, the problem is that

 6     there is a bias in the voters register.  So if people registered for

 7     certain municipalities and were not entitled to, how can I distinguish

 8     between those who were entitled and those who were not but committed an

 9     act of fraud.  And also I see it is unable to pin point and flag these

10     kind of cases.  So this is very deficient source.  This is what I am

11     saying.

12             THE ACCUSED: [Interpretation] I believe that the interpreters are

13     perplexed as well because we didn't receive the interpretation of the

14     last answer, but I would kindly ask the Chamber to arrange for the

15     voters' register from 1996 to be provided to us.

16             JUDGE KWON:  It's not for the Chamber to intervene at the moment.

17     Consult your legal advisor.

18             We'll take a break for half an hour and resume at 3.35.

19                           --- Recess taken at 3.04 p.m.

20                           --- On resuming at 3.38 p.m.

21             JUDGE KWON:  Yes, Mr. Karadzic, please continue.

22             THE ACCUSED: [Interpretation] Thank you, Your Excellency.

23             MR. KARADZIC: [Interpretation]

24        Q.   Dr. Tabeau, when it comes to OSCE, did it annul the 1996

25     elections and did it issue an official report to the effect that the

Page 28293

 1     elections were irregular?

 2        A.   Well, I don't think I have a complete translation, but just to

 3     double-check.  The question is who made this report, what organisation?

 4        Q.   No, no, no.  I will repeat.  We're talking about OSCE, did it

 5     declare the elections irregular or did it at least issue an official

 6     report?  I'm talking about the 1996 elections that were carried out under

 7     the supervision of the OSCE.

 8        A.   Yes, they made their report in which they summarised problems

 9     around the 1996 elections.  So what I'm saying about problems is coming

10     from that report.  There were problems related to the fact that voters

11     could choose for any given municipality, disregarding where they lived

12     before the war in 1991.  So there was no close -- no procedure in the

13     election registration that was checking the place of residence in 1991.

14     It was possible to register for any municipality, and people were abusing

15     this situation.  This is discussed in this OSCE report related to

16     activities in Bosnia and Herzegovina in 1996 and in which there is a

17     separate long chapter about problems around 1996 elections.

18        Q.   If you ask me, there are always problems with elections in the

19     Balkans.  Were the elections declared null and void?  Did OSCE declare

20     the elections irregular and renounce them as such?

21        A.   I am not aware of a statement, official statement, that would be

22     declaring these elections completely invalid, but -- well, I would have

23     to double-check if it was the case.  As of now, I am unaware of such a

24     declaration.

25        Q.   Thank you.  I would appreciate it very much if you would prepare

Page 28294

 1     for tomorrow information as to whether people voted even from America and

 2     elsewhere, if ballots indeed arrived in sacks full of ballots.  Did

 3     people vote wherever they wanted to vote or did they only vote in the

 4     municipalities where they were registered in 1991?  In any case, I'm sure

 5     that you were informed and do you think that the 1997 elections were

 6     general elections for municipal authorities across Bosnia and

 7     Herzegovina?

 8        A.   Yes, that were municipal elections.

 9        Q.   And if I put it to you that the 1997 elections were only for the

10     parliament of Republika Srpska after Mrs. Plavsic walked away from the

11     SDS and decided that there would be a snap election for the parliament of

12     Republika Srpska, did you know that?  Were you informed about that?  Did

13     you check that if you were informed about that?

14        A.   It might be so that that was the case.  What I know is that

15     mainly these elections, 1997, were for municipal parliament.

16        Q.   In all of Bosnia?

17        A.   In the entire Bosnia.

18        Q.   Thank you.  I'm afraid that that wasn't the case.  Do you know

19     that the snap election happened after arguments among the Serbs and that

20     Muslims were not too keen to vote in that election, therefore they didn't

21     register at all?

22        A.   Well, I don't want to comment on the snap elections because I

23     didn't study it in particular.  But from what I can see from the register

24     that I received from OSCE for the 1997 election, it is a complete

25     register in the sense that the entire country is covered, both

Page 28295

 1     Republika Srpska and the Federation of Bosnia and Herzegovina, and all

 2     ethnic groups are represented there.

 3        Q.   Thank you.  Do you know that Muslims and all Muslim parties in

 4     1996 election of 83 MPs in the Assembly of Republika Srpska won 18

 5     places, and a year later only 14 places?

 6        A.   I am not aware of these numbers and I didn't study the results of

 7     elections, you know, especially in 1996.  I don't know that, no.

 8        Q.   But that directly points to the fact that the 1997 list of voters

 9     was much shorter in the Muslim segment than it was in 1996 because if you

10     had to have about 18- to 20.000 people per MP, four MPs less represents

11     about 55.000 to 60.000 Muslim voters who did not vote for their own MPs

12     in that year.

13        A.   Well, sir, what I can tell you about the 1997 voters' register,

14     it comprises 2.5 million individuals from the entire country, and I

15     wasn't particularly interested in the municipalities they voted for.  I

16     was interested in the municipality where they registered to vote.  So it

17     is the physical location of the registration to vote that interested me.

18     Why?  Because I took it as a good approximation of the place of residence

19     at the time of 1997 elections.  So how they voted, for what

20     municipalities, it really didn't interest me at all.

21        Q.   And what if they did not register or if fewer of them registered

22     in 1997 than in 1996, would that have -- be significant for your research

23     because a register or a list of voters is your second bench-mark value

24     with regard to the 1991 census that was compulsory?

25        A.   Yes.  If they did not register to vote, they were not included in

Page 28296

 1     my study.  I only studied those who registered to vote, and there were

 2     2.5 million of registered voters in 1997 elections.  That is one thing.

 3     If you said -- if there were fewer of them registered in 1997 compared to

 4     1996, what kind of problem it creates for me?  It doesn't create any

 5     problem at all because, as I said, 1997 gives more conservative numbers

 6     of displaced persons than 1996, more conservative meaning smaller,

 7     lowered numbers, more conservative numbers of displaced persons.  And

 8     speaking of 1997 election and voters' register versus the 1991 population

 9     census, yes, you are right.  Population census is complete.  It is not a

10     sample.  It is a complete record of the population at the outbreak of the

11     conflict.  1997 voters' register is a sample of this population.  It's a

12     very large sample, 2.5 million individuals out of initially 4.4 million

13     in 1991.  It is such a large sample that in statistical terms this is

14     huge, it is massive.  It is much, much, much more than what I need to

15     produce reliable statistics.  In my report I produced minimum numbers,

16     minimum numbers.  This is at least numbers not complete numbers.

17     Complete numbers are higher than what are presented as the minimum from

18     the sample, from the actually observed data.

19        Q.   Two and a half million adults who had the right to vote in

20     respect of 4 million and 400.000, this obviously surpasses the customary

21     rate of minors by far, does it not, of those who do not have the voting

22     right?

23        A.   Well, you are saying population of age 0 up to 17 versus 18 or

24     older - just double-checking that this is what you mean.

25        Q.   Yes.

Page 28297

 1        A.   Yes, of course.  The 18-plus population, it's much larger than 0

 2     to 17, the minors, indeed.

 3        Q.   What is the specific rate, or rather, what is the natality in

 4     Bosnia and Herzegovina per 1.000 inhabitants?  You did take that into

 5     account, did you not, when you tried to establish the number of those who

 6     were born before 1980?

 7        A.   Well, the rate has nothing to do with distinguishing the

 8     population eligible to vote in 1997 and the rest of the population.  It

 9     is a method of taking those born before 1980 who become 18 in 1997 and

10     are illegible to vote.  I don't work with rates, aggregate rates produced

11     by statistical authorities because I am using a micro data set, micro

12     level data set.  I work with records of information about persons.  Every

13     record represents one individual.  It is a micro level study.  I don't

14     need to use any aggregate level rates.

15        Q.   You covered the whole of the population of those who were born

16     before 1980 and you did not study any patterns, representative pattern,

17     nor did you do any calculations.  You simply applied the list of those

18     who were born before 1980; right?

19        A.   Yes, it is like extracting records and making a long list of

20     those who became eligible to vote in 1997.

21        Q.   Did you take into account everybody, all persons, or did you

22     extract a sample from the totality of the population?

23        A.   Well, I explained already 1991 is a complete population survey.

24     If I take from the census all persons born before 1980, then, as a matter

25     of fact, I take one segment of the population which is complete for the

Page 28298

 1     ages that it includes, and of course it doesn't contain any other persons

 2     that are excluded because of their age.  It is not a sample.  It is a

 3     segment of the population extracted according to the age criterion.

 4        Q.   So you did not need the register of voters.  You could simply

 5     follow those who existed in 1980; right?

 6        A.   If I would just follow them, then what kind of results could I

 7     produce?  I had to look at the voters and cross-reference them with this

 8     segment of the population reported in the 1991 census in order to be able

 9     to study whether they became displaced or not.  So my starting point is a

10     segment of the 1991 census, a segment that is related very strictly to

11     the voters of 1997 election, as the voters most certainly can be found in

12     my segment of the 1991 population.  It's a longitudinal approach.  It is

13     just micro level tracing individuals throughout the time and the sources

14     which report on these individuals.

15        Q.   Thank you.  We'll come back to that when we talk about

16     municipalities.  I want to go back to Sarajevo for a brief moment.  Were

17     you an investigator in the OTP?

18        A.   No, I wasn't an investigator.  The name of my post was

19     demographer.

20        Q.   Very well.  But were you then an investigator working at the post

21     of an OTP, or rather, were you a researcher?

22        A.   I had a post within the OTP, but it wasn't a post of an

23     investigator.  I rather think it was a research position called

24     demographer.

25        Q.   Thank you.  Did you obtain your Ph.D. after that?

Page 28299

 1        A.   After, after I was -- I started working at the OTP, no.  I

 2     obtained my Ph.D. years before.

 3        Q.   Thank you.  Did you know Mr. Robert Donia and did you collaborate

 4     with him?

 5        A.   I know Mr. Donia but we never co-operated.  We occasionally met

 6     during my work at the OTP at the times he was visiting the office.

 7        Q.   Thank you.  If we take Sarajevo as a park, let's use this

 8     metaphor, what kind of fluctuation of population would be significant in

 9     your view that would warrant the application of the capture/recapture

10     formula?  1.000?  2.000?  10.000?

11        A.   Well, it is not about the fluctuation of the population; it is,

12     you know, about the victims, the number of victims.  And when it comes to

13     the reporting of victims in both sources that I used in

14     capture/recapture, there are practically no fluctuations.  That is the

15     nature of the sources as such.  The thing is that having these two

16     sources I restrict my study to the relevant victim records, victim

17     records that died during the siege period as defined in the Karadzic

18     Sarajevo report or the next one and within the front lines in Sarajevo.

19     So there is no fluctuations here other than that my samples are really

20     samples.  I can't be sure that each of these samples is complete.  I know

21     they are incomplete.  They are just samples.

22        Q.   Thank you.  But you take a sample which should represent the

23     whole up to 0.05 per cent margin of error; is that correct?

24        A.   This is -- this error margin you mentioned has nothing to do with

25     sample selection, but, you know, the issue you are mentioning,

Page 28300

 1     fluctuation of the population, would have been relevant if I studied a

 2     different phenomenon, if I studied a disease in a population, and then

 3     there would be a migration out and in, there would be birth, then this is

 4     relevant fluctuations, of course.  But I didn't study diseases, I didn't.

 5     I studied a completely different variable, as you said yourself.  This is

 6     the victim number, that is my variable.  And if we speak of a closed

 7     population as the one within the front lines in Sarajevo, both sources

 8     are not disturbed by any fluctuation.  This is the nature of the sources,

 9     as a matter of fact.

10        Q.   But, Dr. Tabeau, does the number of the fatalities or the

11     casualties can be compared with the number of the population and is it

12     expressed in relative numbers?

13        A.   Yes, you are right.  In the Galic report we calculated the rates

14     as well, and for this a complex procedure was used starting with defining

15     what segment of the Sarajevo population covered by household survey

16     Sarajevo was the original as the one who lived there in 1991.  And

17     approximately 85 per cent of the population covered by the household

18     survey Sarajevo was the original as in 1991 census.  For this we made a

19     pilot study, selected approximately 5.000 questionnaires from a few areas

20     in Sarajevo, questionnaires that were selected disregarding whether or

21     not deaths were reported, just a complete population of 5.000 was picked

22     out.  And based on these questionnaires, we studied the distribution of

23     population in Sarajevo household survey.  As I said, 85 per cent was

24     original 1991 population; 15 per cent was new population, who this or

25     another way came into Sarajevo and stayed there within the front lines at

Page 28301

 1     the time of the survey.  Only the original population was taken for the

 2     calculation of rates, and only the deaths - I mean the killed

 3     people - reported from the original population were taken for the

 4     calculation of the rates.  So it is just narrowing down the material

 5     obtained from Sarajevo household survey, narrowing it down to work with

 6     unbiased information and using this information for the calculation of

 7     rates.

 8        Q.   I would like for us to finish this issue regarding Donia.

 9             THE INTERPRETER:  Could Mr. Karadzic please repeat the number

10     slowly.  Thank you.

11             MR. KARADZIC: [Interpretation]

12        Q.   While we are waiting, can we agree that --

13             THE ACCUSED: [Interpretation] 21434, 65 ter.  This is a book

14     about Sarajevo written by Mr. Donia.  Can we have page 3, please.  We

15     need the paragraph in the middle that begins with:

16             [In English] "According to ARBiH estimates over a million

17     person-passages were made through the tunnel from July the 30th, 1993,

18     until the end of the war."

19             [Interpretation] And then further on he speaks about the

20     transportation of fuel, food, et cetera, et cetera.

21             MR. KARADZIC: [Interpretation]

22        Q.   Did you know that there was such a high frequency of traffic

23     along this tunnel?  And if that is true, one can easily say that actually

24     Sarajevo was not under siege at all?

25        A.   Well, this passage is mentioned, first of all, started in July

Page 28302

 1     30th, 1933 [sic], that is one thing, not in the initial episodes of the

 2     siege.  And secondly, person-passages, what does it mean actually?

 3     Passage?  Where did it end?  Where did they go when they were leaving the

 4     tunnel?  So passages just, you know, for me, one way and back perhaps,

 5     one and back, one and back.  The same person can make thousands of

 6     passages this way and back.  So it doesn't tell me anything about it at

 7     all.

 8        Q.   But did you know that the tunnel was so busy that so many persons

 9     were going to and fro?  It doesn't say exactly, but obviously people

10     travelled in both directions and that tunnel was not part of a closed

11     system; is that right?

12        A.   Well, it says, you know, this passage in the book about

13     person-passages, person-passages, not about persons who were leaving the

14     town through the tunnel and moving out from Sarajevo elsewhere.  If you

15     go down more to the end lines of the same passage, you know, it tells you

16     what was transferred through the tunnel.  So obviously the traffic was

17     two ways.  This is how I read it.  It was busy, indeed, but on the other

18     hand if the Sarajevo was cut off from the outside world, it is obvious

19     that ways had to be invented for providing aid to the people who lived

20     there.  But I don't want to make comments on this because I really didn't

21     study the traffic in the tunnel and who was doing this, how frequently,

22     for what purposes.  Obviously there is a description of it in the book of

23     Mr. Donia, and perhaps this can serve as a source on this.

24        Q.   Thank you.

25             THE ACCUSED: [Interpretation] Can we have these three pages

Page 28303

 1     admitted into evidence?

 2             JUDGE KWON:  Ms. West.

 3             MS. WEST:  I don't have any objection.

 4             JUDGE KWON:  That will be admitted as a Defence exhibit.

 5             THE REGISTRAR:  As Exhibit D2249, Your Honours.

 6             THE ACCUSED: [Interpretation] Thank you.

 7             MR. KARADZIC: [Interpretation]

 8        Q.   We'll come back to Sarajevo.  Please, the research that you did,

 9     was that a scientific research that was designed in advance in terms of

10     specific procedures and sources to be used in the process?

11        A.   Of course it was scientific research in which decisions were made

12     about sources and methods and procedures, and in the course of this

13     research actually many, many activities were taken to guarantee reliable

14     results of it.

15        Q.   I am sorry, it wasn't interpreted.  I asked you whether that was

16     a scientific and empirical research or whether that was research

17     conducted on the basis of the data collected from various sources.  Did

18     you, yourself, design the scientific and empirical research and did you

19     decide the manner in which the data were to be collected and then you

20     processed them statistically; or were you just given the data collected

21     from various sources?

22        A.   I want to stress, nothing is given like that and used like that

23     to produce good, meaningful result.  So in order to propose meaningful

24     research, a lot of effort is done to identify right sources and the ways

25     of processing these sources and produce the results.  So there was a

Page 28304

 1     careful preparation, a very careful preparation of every single report

 2     written for this case and other cases.  That I used sources compiled by

 3     others, like 1991 population census, I can only say I'm extremely happy

 4     the census existed, that professional statisticians prepared the census,

 5     conducted it and made census files, that I am able to use this census the

 6     way I am using it.  I am extremely happy that I have the opportunity to

 7     use voters' registers in the absence of the next census in Bosnia and

 8     Herzegovina.  And for the purpose I'm using the voters' register, this

 9     source is as good as any other professional statistical source.  I am not

10     using voters' register to estimate measures that don't make sense when

11     estimated from this source.  I am using it for a certain purpose and for

12     this purpose it is justified to use this source.  So I see nothing wrong

13     in using information compiled by others, especially if I make sure that

14     the way of compiling this information is professional and acceptable for

15     research purposes.

16        Q.   Thank you.  I'm not attacking you.  Let us just establish the

17     facts.  So you believe that a census is a statistical fact.  Now, as for

18     the rest, are those statistical facts of the same order like voters'

19     registers and other lists, are they or aren't they?

20        A.   Well, census is exceptional among all sources that exist.  So

21     voters is of course a different type of source.  Well, whether it can be

22     compared with the census, obviously not, because the coverage, the

23     content, the way of making the source is different.  But it doesn't mean

24     that we have to disregard the voters' registers as completely useless for

25     our purposes, on the contrary.  I can assure you that these are good

Page 28305

 1     sources for this type of research as we did.

 2        Q.   Well, with all due respect, Dr. Tabeau, we are here adding pears

 3     and oranges.  In your report, you say the reliability is for the most

 4     part satisfactory although it cannot be compared to the usual statistical

 5     sources.  If you made a comparison between the voters' registers from

 6     1996 or 1997, those would be the sources of the similar or identical

 7     order.  But you had to take the data from 1997 when the Muslims were

 8     actually not interested in participating in the elections, and therefore

 9     no comparisons can be made, although you did make a reservation here by

10     saying that they cannot be compared with usual statistical sources.  Now,

11     this causes concern with me.  Why do you say that this is incomparable?

12     Why did you use that term in your report, the term that is now becoming a

13     part of a criminal procedure?

14        A.   Yes, voters' registers are not comparable with the census, as the

15     methodology of making them is not the same as the one of the census.

16     However, voters' registers are nothing less than a sample of census

17     variables.  It is just a sample of basic variables from the census, 1991

18     census, that the OSCE received from the statistical authority.  The

19     variables provided included the names, the date of birth, the place of

20     residence in 1991.  That is what was needed for the OSCE to make a

21     voters' register, a voters' register that could be used by the eligible

22     voters to check whether their names are there.  And that is what was done

23     in Bosnia in 1997 elections.  So voters' register is different from the

24     census.  Census reports on many more variables than just the three I

25     mentioned.  It reports about ethnicity.  It reports about religion.  It

Page 28306

 1     reports about the education, marital status, children.  It reports about

 2     place of birth, about the place of residence, about the profession, about

 3     the employment, about the housing, about the agricultural farms.  It's

 4     huge.  It is unique.  That is the census.  We don't have all these

 5     variables in the voters' register.  Voters' register is a simple register

 6     with names, JMBG, date of birth, place of residence, not good enough to

 7     make a statistical study of the population of Bosnia and Herzegovina in

 8     1997, by all means not good enough to take it as a list of registered

 9     voters who can be compared with the census records and studied from the

10     point of view of their displacement.  These are the differences between

11     these sources.

12        Q.   Thank you.  I am currently disregarding different variables that

13     the register does not comprise.  What I'm primarily interested in is the

14     number.  Do you know that in our country people have to be registered

15     again and again for each election and that in many cases people simply

16     don't register themselves and don't feature in the voters' registers?

17        A.   I'm not sure what is the question, but if it is whether the

18     voters' register is incomplete then the answer is yes, it is incomplete.

19     Only those who registered are there in the register.

20        Q.   Thank you.  Let us just finish the previous question.  Was there

21     any difference between what you did and the data that you obtained or

22     obtained by other sources for various purposes and is it different from

23     research that you set up in advance in terms of standards and procedures

24     that would govern the collection of data and after that you proceeded

25     with the collection of data?  Would that constitute two different ways of

Page 28307

 1     research and which would be your preference between the two?  Would you

 2     prefer to have standards set up in advance?  In other words, that would

 3     be a longitudinal research that would produce the data collected on the

 4     basis of your design.

 5        A.   Well, to keep it short, the question is whether there exists

 6     something like a perfect research design and how far I ended from this

 7     perfect design.  Well, when it comes to the study of displacements,

 8     ideally I would like to use the second census, post-war census.  However,

 9     I don't have it and even today I don't have it.  What I selected, the

10     voters' register, is an excellent replacement of what is unavailable.  So

11     the other alternative that exists is not to do anything at all.  I

12     believe that what has been done with the voters' register and the 1991

13     population census is of high scientific standard and is reliable.

14     Moreover, there are sources that can be compared, independent sources,

15     and results can be confronted.  I did such a comparison and I can tell

16     that this study based on voters' register does not significantly differ

17     from the official registration by UNHCR of displaced persons and

18     refugees.  When it comes to victims, ideally we would all like to see a

19     complete list of all victims with all kind of details and the name of the

20     perpetrator next to each victim.  That is the ideal that we are all

21     looking for.  This ideal doesn't exist.  It's very challenging to produce

22     a list that could be covering a part of this ideal, a pattern that can be

23     produced based on existing sources.  I believe that in all these studies,

24     Sarajevo studies, and in Srebrenica we have come very close to this ideal

25     in terms of numbers, not in terms of perpetrators.  This has never been

Page 28308

 1     the subject of my report.  But in terms of numbers, I think we are very

 2     close and sources that we used are also highly reliable and meaningful.

 3        Q.   Why was then this study performed based on other sources if you

 4     say that you had official registrations made by the UNHCR?  Were there

 5     any flaws in that UNHCR registration that led to it not being applied in

 6     this instance?

 7        A.   Well, I think UNHCR can themselves testify about what they have

 8     done.  So the whole point of this study we made is to make an independent

 9     study based on good sources and compare the results with the work of

10     UNHCR.  So that's the essence of this work.  We work with primary

11     sources, we work with names, we have applied a careful approach I think a

12     good methodology, and this is why it is worthwhile to do this kind of

13     work and to present results as completely independent to what UNHCR has

14     been doing.

15        Q.   Thank you.  Does that mean that the Defence should understand

16     this to mean that the OTP and their services are more objective in their

17     approach to this problem than the UNHCR?

18        A.   Well, first of all, I am an expert for the Prosecutor -- for the

19     Prosecution, but in my opinion I am working to provide the true picture

20     of what happened.  And I have been doing this the best way I can by

21     studying always all ethnicities, complete areas, complete time episodes.

22     I work with many sources.  I cross-reference the sources.  I confront the

23     sources with each other in order to present as independent picture as

24     possible.  I don't consider myself working for the Prosecution or for

25     anybody else.  I am -- my role is to provide an independent opinion on

Page 28309

 1     what happened.  It happened that I was part of the OTP 11 years long;

 2     that is correct.

 3        Q.   [Microphone not activated]

 4             THE INTERPRETER:  Microphone, please.

 5             JUDGE KWON:  Just a second.  Yes, could you repeat.

 6             MR. KARADZIC: [Interpretation]

 7        Q.   Maybe we shall -- back to this issue later on.  Let me ask you

 8     this:  There is no dispute between us that the bureau of statistics is

 9     fully qualified and competent to carry out census.  Now, as for the

10     sources who collected the data that you used, are these sources -- have

11     equally valid credentials?  Are they certified?

12        A.   Well, I don't know exactly what you mean, "certified,"

13     credentials.  I used the ICRC list of missing persons, ICMP list of

14     identifications, Sarajevo household survey.  I think these are -- yeah,

15     international organisations, very well-known, and who operating -- has

16     been operating for very long time, especially ICRC, and I believe this is

17     the best credentials as I can have for these sources.  Yeah, I also use

18     Sarajevo household survey, a survey that was conducted by a local -- by

19     group of local researchers in Sarajevo.  And I actually checked the

20     questionnaire they developed and the procedures they used for completing

21     the survey, but the computerisation of these questionnaires and data

22     processing, cleaning of the data, everything was done here by my unit and

23     others who were involved in projects with me.  So even though the

24     questionnaire was developed down there, in Sarajevo, first I see this is

25     a good questionnaire; second, I received the original questionnaires for

Page 28310

 1     my processing, I arrange everything myself according to good standards.

 2     So I don't think we have a problem here as well.

 3        Q.   Perhaps I wasn't clear enough.  I am not dealing with Sarajevo

 4     only.  I'm asking you in principle.  For instance, if, for example, a

 5     music society were to come up and give you their information and when we

 6     use the word "music society" or singing society, that means that there's

 7     nothing binding involved.  For example, who is this, PIP, the

 8     Podrinje Identification Project, are they certified for working on

 9     statistical methods?  Are they competent?  Are they objective?  And has

10     that been confirmed anywhere?

11        A.   That is Podrinje Identification Project that you mentioned is not

12     a source I used.  It is just an organisation down there, in Bosnia, who

13     works together with ICMP on the identification of victims.  What matters

14     for me is the ICMP itself, International Commission For Missing Persons,

15     who are certified, I am sure, and there was probably another person from

16     ICMP to tell about the certificates they have.  But for me what mattered,

17     I had the occasion to visit ICMP several times in the course of my work

18     at the OTP.  I had the opportunity to check personally how they work,

19     what kind of procedures are in place.  For the whole process from, you

20     know, taking samples from the remains, sending them to the lab, DNA

21     profiling, you know, putting into the database, the same with the blood

22     samples of the relatives, how the matching is done, how the coding is

23     done, how the storage is done.  So I could have assessed this for myself,

24     and that is, in my view, the best certificate in my eyes, you know, about

25     the reliability of this source.  But of course the methodology ICMP uses

Page 28311

 1     is a worldwide standard, so it is not that they invented the method of

 2     DNA profiling and matching; they just use the methodology.  And I use the

 3     results of the matching.

 4        Q.   I was not referring to the ICMP, although I think that at that

 5     time even the ICMP hadn't had an international licence.  However, we had

 6     their witnesses here as well.  This is what I'm interested in, though,

 7     that is the extent to which you relied on sources.  Now let us look at

 8     something else.  As far as municipalities are concerned, you did some

 9     research on demographic trends in relation to the consequences of war, or

10     rather, causes and consequences of war in 27 municipalities and then

11     especially within that context in nine municipalities.  Do you know how

12     many municipalities exist in Republika Srpska or do you accept, rather,

13     that there are 62 of them?

14        A.   Well, I don't remember the exact numbers, but if you are telling

15     me 62, I accept this number.  I will check it, though.

16        Q.   Thank you.  So is it the task of the Prosecution or was it the

17     task of the Prosecution - and within that context your task as well - to

18     provide evidence on lawful or unlawful action taken by the Serb

19     side - and myself in particular - in relation to a certain area?

20        A.   Sir, I said earlier in my testimony that I didn't study causes of

21     migration, and of course I didn't study all lawful or unlawful actions of

22     yourself or anybody else.  It is not part of my report.

23        Q.   Thank you.  But this is what I'm interested in now.  What do

24     these 27 municipalities mean?  Are they a sample for all of

25     Republika Srpska?  Or even more, for the entire area that we had under

Page 28312

 1     our control during the war?

 2        A.   The 27 municipalities were selected according to the indictment

 3     of this case.  This is initially as it was presented at some point, at

 4     the time the report was made.  So it is not a sample.  It is just the

 5     area covered by the indictment of the case, Karadzic case, that is

 6     studied in our municipalities report.

 7        Q.   Thank you.  But you expressed all results in relative terms as

 8     well, didn't you?

 9        A.   As percentages, you mean.  That the change in the ethnic

10     composition was expressed in percentages.  But still, you know, I used

11     both absolute terms and relative terms in relation to the 27

12     municipalities selected in accordance to the indictment of the Karadzic

13     case.

14        Q.   Since you're a statistician and a demographer, would you have had

15     a different result if you added to the 27 municipalities a 28th one, for

16     instance, Gradiska and as a 29th one Srbac, and as a 30th one Derventa,

17     say all of those municipalities, then the entire autonomous region of the

18     Krajina that had 18 or more municipalities itself.  What would the result

19     be given that sample, or, rather, given this mass that would be looked at

20     statistically?

21        A.   Well, the numbers that I present relate to the 27 municipalities.

22     If I change this number, if I add new municipalities, the numbers must

23     change obviously because the calculation will be done for a new area that

24     is different from the 27 municipalities.

25        Q.   Thank you.  Out of these 27 municipalities, if we were to take

Page 28313

 1     nine municipalities where the fiercest fighting took place and where the

 2     denouement was the most dramatic then these results become even more

 3     dense, isn't that right, I mean they're even more convincing as claimed

 4     by the indictment?

 5        A.   I think that it doesn't work like this.  You know, if we ask the

 6     question how many persons became displaced from the area of 27

 7     municipalities versus nine out of 27, so that obviously for nine

 8     municipalities the number will be smaller than for 27.  That is as simple

 9     as that.  Ethnic composition might be then showing more dramatic changes

10     for the nine municipalities, of course.  If these were -- municipalities

11     were -- these events, as you call them were dense.  But it is not that we

12     were making a selection in our report.  We make a report consistent with

13     the indictment area, and there is no manipulation here with the number of

14     municipalities.  The indictment says 27 municipalities.  We take 27 and

15     make an analysis for the 27.  And the other day last week there was a

16     mention that seven have been dropped out of 27, so it would be good to

17     see the picture for the 20 municipalities instead.  That is how it works.

18     What comes out is just the result of the analysis.

19        Q.   Thank you.

20             THE ACCUSED: [Interpretation] Could we please call up in e-court

21     chapter 3 are the main findings.  They have to do with Croats --

22     actually, it's all the same to me.  Any map of these municipalities, say

23     number 5 from this chapter.  I'm calling this up not because of the

24     Croats but because of the image itself.  For instance, we can also have

25     number 7 from chapter 4.  Any one of them.  We just need to see the map.

Page 28314

 1             MR. KARADZIC: [Interpretation]

 2        Q.   You are certainly informed that I was president in all of

 3     Republika Srpska; right?

 4             THE ACCUSED: [Interpretation] P4994, page 5 within this number.

 5             MR. KARADZIC: [Interpretation]

 6        Q.   Do we agree that I was president - I mean that I was

 7     president - in all of these municipalities, even in those that in Dayton

 8     we handed over to the Federation; right?

 9        A.   Yes.

10             THE ACCUSED: [Interpretation] Can we have the coloured one

11     enlarged.  We don't really need the Serbian version.

12             MR. KARADZIC: [Interpretation]

13        Q.   Dr. Tabeau, since I have been accused of systematic persecution,

14     a system that was applied in Republika Srpska, do you believe -- for

15     example, look at the west, Dubica, it is north of Prijedor, 009; then

16     Gradiska, also to the north of Prijedor; then Prnjavor, then Srbac, these

17     are just four municipalities closer to the Sava River up there.  Had you

18     included them, what would this system of persecution look like

19     percentage-wise?

20        A.   Well, I don't know that, but if you want and you ask me then I

21     can recalculate some statistics for you for the entire Republika Srpska.

22     But I don't think that entire Republika Srpska is part of the indictment.

23        Q.   But if it's a system that is involved, if I have been accused of

24     systematic behaviour based on a system, does it not mean that in some

25     municipalities they did listen to me and in others they did not?  And did

Page 28315

 1     they obey me in the municipalities where there hadn't been any

 2     persecution, or did they obey me in the municipalities where there hadn't

 3     been any persecution?

 4             MS. WEST:  Objection.

 5             JUDGE KWON:  It's not for the witness to answer that question.

 6             THE WITNESS:  Well, I just want to comment on the figures for

 7     Republika Srpska.  They are available from the report, municipalities

 8     report.  On page 66, for example, in table 2BH, there is an overview of

 9     the minimum number of displaced persons for the entire country Bosnia and

10     Herzegovina with the division into Republika Srpska and the Federation.

11     And there is a similar table for the estimated number of displaced

12     persons and refugees which is next page, 67, and there is the change in

13     the ethnic composition on page 65.  So pages 65, 66, 67 contain these

14     statistics.  So I'm not hiding anything.  Whatever I have, I present in

15     this report.  So there is a context for the figures for 27 municipalities

16     in terms of the figures for Republika Srpska as a whole.  Well, I don't

17     think the picture is so dramatically different really.  The numbers of

18     displaced persons are larger, though, and very much so.

19             MR. KARADZIC: [Interpretation]

20        Q.   But you certainly do not deal with displaced persons only.  You

21     also deal with the density of the events concerned, and if we look at the

22     framework involved it is of crucial importance to look at things in their

23     entirety.  One of the key values involved is the total area because it is

24     expressed in real terms.

25        A.   I can only say it's all there, it's all there, the 27

Page 28316

 1     municipalities, the entire Republika Srpska, it's all there in the

 2     report.

 3        Q.   Do you believe that establishing the density of events in nine

 4     municipalities in 27 municipalities is insignificant from the point of

 5     view of this criminal case and from the point of view of the indictment

 6     for the purposes of which you worked?  Is the density of the developments

 7     concerned of significance?

 8        A.   Well, I don't know exactly what do you mean by "density," in

 9     terms of relative measures certain percentages are presented in the

10     municipalities report.  That's the only relative measure that is

11     discussed there.  So I think this is the density, I guess, you referred

12     to.  And if I have to answer whether it is right to show these densities,

13     meaning the fractions of displaced persons in various configurations, the

14     answer is, yes, it is 100 per cent correct to calculate such percentages

15     for every municipality in Bosnia and Herzegovina for the entire area of

16     political entities, Republika Srpska and Federation, and for any other

17     area composed of selected municipalities.  It all depends what is -- why

18     these statistics are needed.  In this case, the calculation is presented

19     for 27 municipalities because this is how the indictment area was

20     defined.  If the area would be larger, like the entire Republika Srpska,

21     we would be including consequences that are not necessarily -- that would

22     be not covered by the indictment, which would be completely unjustified

23     then.

24        Q.   Thank you.  Please focus on chapter 2 of your report with the

25     annex for this case, the main findings that pertain to Muslims, so I'm

Page 28317

 1     looking at 2.1.  And you say in 27 municipalities included in the

 2     indictment the share of Muslims decreased from 41 per cent, as it was in

 3     1991, to 32.8 per cent in 1997, or, rather, the decrease was 19.9

 4     per cent.  First of all, do we agree that the share of Muslims in the

 5     general population does not depend only on trends involving the Muslims

 6     but also on trends involving others, the arrival of Serbs in this area,

 7     does it decrease the percentage of Muslims that are there?

 8             MS. WEST:  Mr. President, this is e-court page 9.

 9             THE WITNESS:  Yes.  Now to the question about the trend.  Well,

10     there is no trend in my report.  There is this one 1991 population as

11     reported in the census and ethnicity is reported as of the moment of the

12     census, and it is from -- coming from an open-ended question, so

13     everybody tells what he or she thinks he or she is, end of story.  There

14     is no trend.  So the ethnic composition is calculated on this given

15     moment of time.  And the same is done again for 1997.  The whole point is

16     about, you know, movements of the population, that certain groups moved

17     out, certain other groups moved in in a given territory.  So if the

18     ethnic composition is shown on two different moments of time, then there

19     is a change, a change which has nothing to do with any trend in a report

20     being of ethnic or anything like that because still the same definition

21     of ethnicity is used.  It is just the two -- these two ethnic

22     compositions summarise the change that took place between the two moments

23     of time.  And if one group is gone and replaced by another group, this

24     will be seen in the change of the ethnic composition.

25        Q.   However, if one group hadn't left and the other group increased

Page 28318

 1     its numbers, does that affect the structure?

 2        A.   Of course this will be reflected in the structure.

 3        Q.   Thank you.  So 41 per cent and 32.8 per cent is the share of the

 4     Muslims in the general population in these 27 municipalities; right?

 5        A.   Right.

 6        Q.   Thank you.  Then taking into account all the things that can

 7     affect this, including Serbs moving in, for instance, we cannot just take

 8     it for granted that 19.9 per cent Muslims left; rather, perhaps more

 9     Serbs moved in.  Right?

10        A.   Well, I think the beauty of this report is said that you have

11     numbers of the population movement for every ethnic group.  So these are

12     the displacements, for Muslims, and Serbs, Croats, others.  So on one

13     hand there is this change in the ethnic composition, but on the other

14     hand you just go to figures related to the displacement which are

15     discussed in paragraph 2.2 and 2.3 in the same section.

16        Q.   Thank you.

17        A.   So for my ...

18        Q.   We'll get to that, Ms. Tabeau.  We'll get to that.  So does this

19     mean the following:  That in these 27 municipalities in 1997 there were

20     67.2 per cent non-Muslims, that is to say Serbs and Croats, if we're only

21     looking at point 2.1?

22        A.   Yes.

23        Q.   Thank you.  How did you calculate the 19.9 per cent, on the basis

24     of absolute numbers or did you subtract 32.8 from 41 per cent and then

25     did you calculate the percentages of percentages?

Page 28319

 1        A.   It's the relative difference, this is how it is called in

 2     statistics, it's the difference between 41 per cent and 32.8 per cent

 3     divided by 41 per cent.  It is a relative difference, a relative

 4     difference.

 5        Q.   Thank you.  Would it -- wouldn't it be more accurate to use

 6     absolute numbers and then to base the percentage on the difference in

 7     absolute numbers?

 8        A.   Well, we don't do this with absolute numbers because for 1997 it

 9     is a large sample but a sample.  So if we look at the absolute numbers,

10     this could be misleading.  Some people could take the absolute numbers as

11     the size of the population in 1997, which is not true.  It is just a

12     sample.  So it is much more correct to work with just percentages.

13        Q.   But please take a look at this.  41 per cent and 32.8 per cent

14     are percentages that have to do with the general population, whereas

15     19.9 per cent are percentages within the Muslim community itself; right?

16        A.   Well, it is a change in the fraction of Muslims between 1991 and

17     1997.  That change expressed as a percentage, it is almost a 20 per cent

18     decline, almost 20 per cent.  That's it.  I -- what is the problem here?

19     I really don't understand.  It would be of course easier to look at 1.000

20     versus 2.000 or 10.000 versus 50.000.  Of course it would be, but we

21     don't have complete numbers for the 1997 population so we have to live

22     with percentages.

23        Q.   Thank you.  However, I'm trying to say the following:  When Serbs

24     move into Serb areas, that could also change the share of Muslims in the

25     general population; right?  And we cannot say there are less Muslims in

Page 28320

 1     the general population by 19.9 per cent?

 2        A.   The 19.9 per cent is the change in the ethnic composition, in the

 3     fraction of Muslims, in the fraction of Muslims in 1997 as compared with

 4     1991.  And in terms of absolute numbers, we don't know how many Muslims

 5     left in 1997 in these 27 municipalities.  We can only tell a minimum

 6     number of Muslims living in these 27 municipalities.  That's all we can

 7     tell.

 8             JUDGE KWON:  Mr. Karadzic, it's time to adjourn for today, but

 9     before doing so just shall we do the calculation here.

10             Doctor, 41.0 minus 32.8 is 8.2.

11             THE WITNESS:  Yes.

12             JUDGE KWON:  That should be exactly 20 per cent, why 19.9

13     per cent?

14             THE WITNESS:  I don't know.  I have to recalculate if it is not

15     exactly as Your Honour thinks it should be, yeah.

16             JUDGE KWON:  Very well.

17             Yes, Ms. West.

18             MS. WEST:  May I just quickly do some corrections to the

19     transcript, Your Honour, please.

20             JUDGE KWON:  Yes.

21             MS. WEST:  Page 60, line 12, there are two times where the

22     witness said DEM-2, D-E-M-2.  On the transcript, it's too, t-o-o.

23     Page 62, line 3, there is a blank.  It should be JMBG, J-M-B-G.  Page 63,

24     line 24, the words should be "deficient," not "definition."  Page 55,

25     line 2, it should be "counts of victims," not "kinds of victims."  And

Page 28321

 1     then page 77, the witness was talking about that the voter registration

 2     was not good enough to make a study in 1997, but in line 20 she said by

 3     all means good enough to take it as a list of registered of voters, if

 4     there's any disagreement on here.

 5             JUDGE KWON:  Thank you, Ms. West.

 6             We continue tomorrow morning at 9.00.  The hearing is now

 7     adjourned.

 8                           --- Whereupon the hearing adjourned at 5.01 p.m.,

 9                           to be reconvened on Wednesday, the 2nd day of

10                           May, 2012, at 9.00 a.m.