by Felix Hansen and Sebastian Schneider
The NSU trial has been running in Room A101 of the Munich Criminal Court for four and a half years now. After 374 days of evidentiary hearings,  the closing arguments are being heard now, and the final judgment is moving closer. All that is missing – if everything should go as planned – are the closing arguments of the public prosecutor, the accessory prosecution,  the defense lawyers and any final words by the defendants. This monumental trial will go down in German history. We have prepared statistics about some aspects of the NSU trial. In addition to figures, it shows what was examined in the trial – and what was left out. We have evaluated and mapped mostly raw data, kindly provided to us by accessory prosecutor Antonia von der Behrens. The figures are based on a list maintained throughout the entire trial.
One tenet of German criminal proceedings is the “principle of orality” (Mündlichkeitsprinzip). This means a judgment may only be based on material presented or explained orally at the main hearings. There are only few exceptions to this principle, such as the “self-reading procedure” (Selbstleseverfahren), where the participants read documents on their own outside the court, without having them read out loud in court. Expert and witness testimonies play a central role in criminal proceedings.
540 witnesses and 56 experts testified
By the end of the evidentiary hearings, 596 individuals had been heard in the NSU trial, consisting of 540 regular witnesses and 56 experts. Some witnesses and experts were summoned and heard repeatedly; 774 summonses in total were served. Sometimes individuals were heard repeatedly on the same evidentiary issue, sometimes on multiple issues. Especially police officers and experts were heard on multiple issues. For example, the forensic expert Oliver Peschel from Munich was heard 9 times; Jeanette Pf., police officer in a special task force (“EG Trio”) of the political crime department (Staatsschutz) of the Federal Criminal Police Office (Bundeskriminalamt) was heard 8 times. Permanent witnesses are not included in the figure of 774 summonses. This applies especially to psychiatric expert Prof. Dr. Saß, who created a forensic-psychiatric assessment of Beate Zschäpe, for which he had to attend a large number of sessions to observe the evidence and Zschäpe’s demeanor throughout the main hearings. However, only 12 trial dates towards the end of the evidentiary hearings were scheduled for Saß to actually present his assessment and answer questions. There were 143 follow-up summonses of regular witnesses and 35 follow-up summonses of experts. 
Who requested the summonses?
One telling indicator is who initiated the summonses of witnesses and experts. The senate is required by law to summon all witnesses and experts whose knowledge it considers relevant to the proceedings. So technically all summonses are issued by the senate. The trial participants do have the right to summon their own witnesses, but this right was exercised only twice throughout the NSU trial. The question is therefore, who initiated the summonses. Were the witnesses summoned because they are mentioned in the indictment of the public prosecutor or in the attached list of witnesses, which was compiled before the main hearings, because the senate identified the need to hear them? Or were they summoned because the defense or accessory lawyers requested that witnesses not mentioned in the indictment be heard, and the senate approved these motions? 
Who was heard?
The evaluation shows that most of the witnesses and experts (333) were summoned after being named in the indictment. The public prosecutor submitted not a single formal or informal application to summon witnesses during the main hearings. The senate proactively summoned 190 individuals without any formal or informal application by other trial parties and 33 individuals upon application of the accessory prosecution. Ralf Wohlleben’s lawyers were particularly active; 37 witnesses were summoned at their request, while Zschäpe’s lawyers summoned only 3. The lawyers of the other three defendants showed no activity in this field, and no witnesses were summoned at their request. These figures alone disprove arguments brought in the media throughout the first half of the trial, which accused the accessory prosecution of protracting the trial by being “too active.” 
13 secret service officers and 8 informers testified
13 members of the domestic secret service (Verfassungsschutz) were heard during the trial. Most belonged to the responsible state-level offices of Thuringia and Hesse (due to the Andreas Temme complex). 6 informers who had formally signed undertakings (V-Leute) and 2 who had not (Gewährspersonen), including one woman, also testified. These include only the witnesses whose activities as informers have been verified. It cannot be excluded that other informers were heard, who are still undercover. One example is Stephan “Pinocchio” Lange, who was summoned in April 2015. Only much later did it transpire that this former head of the German Blood & Honour chapter had been an informer for the Verfassungsschutz (news report [DE]). Since he was still undercover at the time of his hearing, his informer activities were not an issue, so we did not include him in the count of summoned informers.
4 of the 8 informers worked for the Verfassungsschutz of Thuringia; the other 4 reported to the criminal police authority (LKA) of Berlin and the Verfassungsschutz offices of Brandenburg, Bavaria and Hesse, respectively.  Most of these individuals were not included in the public prosecutor’s list of possible witnesses. The public prosecutor was obviously interested in one single informer only:  Tino Brandt, whose informer reports were indispensable to the prosecution and who was heard on 5 trial days. More telling, however, are those informers who were not heard. Although the accessory prosecution requested him to be summoned, the Verfassungsschutz informer “Tarif” (Michael Doleisch von Dolsperg, born See) was not heard (see informer profile for Michael See [DE]). Ralf Marschner, code name “Primus,” was also not summoned. There are serious clues, as confirmed by the 2nd NSU investigative committee of the German parliament, that NSU member Uwe Mundlos worked for Marschner’s company in Zwickau. (For info on the investigative committee, see 1 or 2.) This shows how narrow the evidence considered in the NSU trial was. In both cases, the senate did not find it necessary to summon the informers. The public prosecutor even opposed the accessory prosecution’s applications to hear the witnesses.
Blatant omissions in the investigations
The fact that Marschner was not heard shows another obvious imbalance: The supporters in Chemnitz received more attention than others. 23 individuals, including 5 women, from the Chemnitz neo-Nazi scene were heard. 9 had provided verifiable support to the NSU core trio Böhnhardt, Mundlos and Zschäpe after they went underground. 4 of these supporters refused to testify in court, but had already provided testimony to the police. Therefore, a lengthy procedure was necessitated, in which various interrogators had to reiterate the supporters’ previous testimonies.
The NSU core trio spent its early underground phase in Chemnitz, and quite a lot is already known about this time. In 2000, the three fugitives relocated to Zwickau, which lies 50 km away. Therefore the question who supported the NSU in Zwickau – apart from the defendant André Eminger and his wife, who already knew the core trio from Chemnitz – remains largely undealt with. This indicates blatant omissions in the investigations of the public prosecutor and the Federal Criminal Police (Bundeskriminalamt). Obviously they quickly settled on the presumption that the NSU was an isolated trio with few supporters, who had known the group from Jena and Chemnitz, before they became fugitives. By refusing to summon Ralf Marschner, the senate also displayed its lack of interest in further investigating the time in Zwickau.
The senate did, however, look closer at the neighbors of the NSU core trio: 20 neighbors were heard, one from Chemnitz, 9 from Polenzstraße in Zwickau and 10 from the vicinity of the NSU’s later residence at Frühlingsstraße. Some of these had also witnessed when Zschäpe blew up the building on 4 November 2011.
Evaluating the gender ratio of witnesses and experts shows that 16% among the groups “Supporters” and “Nazis” (3 and 7, respectively) were women. A female-to-male ratio of about 1/5 is typical for the Nazi scene (see “Das Fünftel der Szenen” [DE] by Eike Sanders).
14% of the heard police officers were women and only 7% of the experts.
NSU murders and terror attacks in court
First NSU attack hardly treated in trial
310 individuals were summoned to testify on the crimes of the NSU: murders, terror attacks and robberies. Witnesses include accessory plaintiffs, witnesses and investigating officers.  It is no surprise that the highest number of witnesses and experts (53) testified on the bomb attack at Keupstraße in Cologne. The nail bomb was designed to kill and injure as many people as possible in a neighborhood with many migrants. For this attack, the indictment accuses Beate Zschäpe of complicity in 22 attempted murders in legal unity with 22 cases auf causing bodily harm by dangerous means. The senate even raised this figure and at the end of the evidentiary hearings indicated that 32 individuals were now considered victims of attempted murder, with 23 cases being in unity with causing bodily harm by dangerous means. But this extended indictment still fails to reflect the dimensions of this act. It is a miracle that this bomb did not kill anyone, since it was filled with around 800 extra-long nails. But the “attack after the attack” – the racist police investigations, the public allegation that the attack was related to organized crime, the spreading of suspicion among neighbors – are also not covered by this indictment.
The comparatively high number of witnesses heard on the murder of Halit Yozgat is not just due to the higher number of witnesses present at Yozgat’s internet café, but also to the fact that the Verfassungsschutz officer Andreas Temme was one of them. His role remains unclear. Temme alone was questioned on 6 trial dates. Other officers of the Hessian Verfassungsschutz and Temme’s wife also testified. It is clear to all attentive trial observers that Temme’s story cannot be true the way he tells it. Nevertheless, after 18 sessions of investigating the Temme complex, the senate reached the rather exclusive conclusion that Temme had “objectively, comprehensibly and plausibly [described]his observations in connection with the crime of 6 April 2006 in Kassel.”
The bomb attack with a pipe bomb camouflaged as a flash light in a bar in Nuremberg is a special case, because it was not revealed by the investigations, but by the testimony of Carsten Schultze in the main hearings. This attack is therefore not part of the indictment and was only introduced through the questioning of a police officer who had led the investigations. After Schultze’s testimony, the public prosecutor initiated investigations into this attack, only to later stop them, because they would not have contributed significantly to Zschäpe’s expected sentence.
Murders were processed quickly
Some issues were given a lot of space in the trial, while other complexes were handled with irritating speed. Also owing to the lack of witnesses, the murders were processed very quickly. One example for how little time was invested to solve individual cases is the murder of Mehmet Turgut in Rostock on 25 February 2004. The circumstances of this murder were examined on 4 trial dates, where 8 witnesses were heard. 4 police officers and only 4 other witnesses were heard: One medical expert, one local resident who had heard the shots, the man who owned the snack stall where Turgut was murdered and had found his dead body, and one witness who was the last to see Turgut alive. Only one trial date was dedicated to the crime; the other three days dealt with other issues. More than some of the other cases, the murder of Mehmet Turgut largely remains a mystery, especially concerning the choice of victim and possible local support for the killers. Relatives of the victim were not questioned in court about the aftermath – even though Mehmet Turgut’s brother regularly attended the trial in Munich as accessory plaintiff. Overall, only members of 4 families of victims were heard as accessory plaintiffs and had the chance to speak about the effects of these crimes. Nevertheless, some public voices still proclaimed that the victims’ families were given too much space in the trial.
Which issues were witnesses heard on?
The turning loop complex
The “turning loop complex” is an example of how much space one individual set of evidence can take up. The defendant Carsten Schultze already testified on the 6th and 8th trial date about an assault by neo-Nazis against two individuals at the last stop of the tramway in Jena-Winzerla. Schultze admitted to kicking one of the two individuals in the back and further reported that Ralf Wohlleben had been party to the assault. In his defense statement two years later, Wohlleben contradicted Schultze’s testimony, claiming said incident had never occurred. Wohlleben’s defense later requested that members from the neo-Nazi scene during this period be summoned, whom Schultze had named as parties to the assault.
The strategy of Wohlleben’s defense was not to answer the question whether he had been party to the assault, but rather to weaken the credibility of Carsten Schultze. Schultze’s testimony about the procurement of the Ceska murder weapon had severely incriminated Wohlleben. So by attacking Schultze’s testimony about the assault, Wohlleben’s defense lawyers aimed to undermine Schultze’s general credibility. They even referred to Schultze as a “crown witness.”
It is no surprise that the Nazi witnesses summoned at the request of Wohlleben’s defense claimed not to or hardly to remember said “brawl.” Wohlleben’s defense went to great lengths and even tried to summon the local mayor of Jena-Winzerla, two employees of Jena’s public transport association and the former owner of a gardening company. Issues they requested to consider were the structure of the turning loop at the tramway terminus and whether there used to be a wooden hut at the site, as described by Schultze.
But Wohlleben’s defense had obviously not expected the accessory prosecution’s meticulous and dedicated research. Attorney Hardy Langer did what the public prosecutor and the Criminal Police should have done: He searched the local papers at several libraries and found a matching report in the Ostthüringer Zeitung. This report allowed him to find the victims of the attack described by Schultze. Both were heard in the trial and confirmed Schultze’s testimony. So while Wohlleben’s defense had unraveled the case to undermine Schultze, it turned out to reinforce his credibility. Accessory prosecutor Eberhard Reinecke fittingly described this as a “shit goal” by Wohlleben’s defense. 11 witnesses were heard on this set of evidence alone; aerial photographs were examined and the newspaper article was read out. This assault was a recurring issue in the NSU trial in 2013 and 2015, and more intensively from July to December 2016.
248 applications for evidence
Many applications for evidence by accessory prosecution
The number of applications to take evidence (or suggestions to examine evidence) also sheds a light on the trial.  It is telling that a large majority of applications (152) were submitted by the accessory prosecution. Since the accessory plaintiffs desired a comprehensive investigation, their lawyers aimed to introduce certain issues into the trial, which were neglected or simply ignored by the public prosecutor. After all, the NSU trial is the only official option for accessory plaintiffs to actively participate in the investigation. In some cases, the accessory prosecution successfully introduced issues (such as the Temme complex or the role of informer Carsten Szczepanski).
The evidence introduced by the accessory prosecution also provided support to the indictment (such as the NSU’s spying out of the synagogue at Rykestraße in Berlin, the turning loop complex, the question of Zschäpe’s participation in editing the confession video, Zschäpe’s attendance at the “1997 Hetendorf Conference Week” (Hetendorfer Tagungswoche; a neo-Nazi meeting organized by the organization Artgemeinschaft), the complex of André Eminger’s “White Brotherhood Erzgebirge” and the “Aryan Law and Order” fanzine, as well as the Sczepanski complex.) Unsurprisingly, the public prosecution based its closing argument on issues which would never have entered the trial without the accessory prosecution’s efforts. This alone dismantles the legend that the accessory prosecution strategically protracted the trial. Anyway, the senate refused to approve most applications by the accessory prosecution, no matter how well-founded (as in the case of Ralf Marschner). Investigations into important events, such as the shredding of files by the Verfassungsschutz or about possible NSU supporters in Dortmund, where Mehmet Kubaşık was murdered, were rejected by the senate. Therefore, only 33 witnesses requested by the accessory prosecution were summoned, despite the much larger number of applications.
Sluggish “questioning” of Zschäpe
A lot of trial time was wasted on Zschäpe’s recurring and partially successful attempts at replacing her defense lawyers and the parallel attempts by her original lawyers to be released from their mandates. A great deal of time was also dedicated to Zschäpe’s cumbersome testimony method. Starting in late 2015, she was “questioned” by means of written-out notes, which were first read out and then answered in writing and again read out by Zschäpe’s new lawyers several weeks later. A psychiatric assessment of Zschäpe and two counter-assessments also occupied the court for almost half a year.
The motions for judicial disqualification are also interesting in this respect. 26 motions to disqualify the senate or individual judges were made: 17 by Wohlleben and 9 by Zschäpe.  We counted approx. 25 fully canceled trial dates due to disqualification motions against the judges. A few disqualification motions were also made against expert witnesses, one of them by the accessory prosecution. This particular one against the expert Prof. Bauer, chosen by Zschäpe’s new defense, was the only one to be granted.
Our last chart maps the sluggish progress, particularly since 2015. It shows how many dates were actually spent on hearings and how many experts and/or witnesses were heard on these dates. In the spring and summer of 2015, the trial dates were cut from three to two sessions per week out of concern for Zschäpe’s health. This changed after she received a new lawyer, but from the summer of 2015 more and more trial dates were canceled on short notice by the court or for other reasons. Our evaluation shows that the number of summoned experts and witnesses also began to decline around this time.
The NSU trial has reached gigantic dimensions: The evidentiary hearings alone lasted more than four years; almost 600 individuals were heard in court and approximately 250 applications to take evidence were made. But the analysis shows that some areas we find to be of extraordinary importance were hardly treated in the trial or not at all. Some of the NSU’s crimes were dealt with very briefly. Questions of serious concern to many relatives, such as how the victims were chosen and whether there were any local supporters near the crime scenes, were barely asked and remain unanswered. The question about local supporters of the NSU core trio in Zwickau was bypassed almost completely. Likewise, the court and the public prosecutor blocked nearly all attempts to investigate informers in the NSU support network or internal procedures, such as the shredding of files about informers by the Verfassungsschutz of Thuringia one week after the discovery of the NSU.
The evaluation also clearly disproves the accusation against the accessory prosecution of protracting the trial by demanding a full investigation, which was raised by certain news media. Instead, the trial owes its duration to the lengthy conflicts surrounding Beate Zschäpe’s defense and her unconventional method of testifying in writing. The senate needed so long to question Zschäpe, that the accessory prosecution had to wait until July 2016 to ask its questions – which Zschäpe refused to answer. The situation was aggravated by the senate’s irregular scheduling of sessions, especially after the summer of 2015.
Despite all the figures, many facts of the trial cannot be quantified. In many of the sessions, the cruel effects of the murders and terror plots on the victims and their families could be felt. The trial gathered enormous amounts of material and knowledge, not just from the actual testimonies, but thanks to the well-researched evidentiary applications by the accessory prosecution. This data founds the basis for any research into the NSU complex beyond the trial.
The evaluation is based on a list continuously maintained since the beginning of the trial by accessory prosecutor Antonia von der Behrens to the best of her knowledge and ability. Although this list has been reviewed several times, inaccuracies cannot be fully excluded due to the number of trial dates, witnesses and applications. Therefore we cannot guarantee the correctness of every individual figure. However, we do assume the data’s high quality. We have chosen particular categories to classify evidence and witnesses, but of course this does not exclude other classifications. The text points this out at the corresponding locations.
 For more info see: https://en.wikipedia.org/wiki/Public_Prosecutor_General_(Germany)
 Some media also mentioned figures for witnesses and experts which deviated considerably from ours. For example a video by “Spiegel-Online” [DE] mentioned 815 witnesses and 42 experts. These figures are probably taken from information by the press agency DPA [DE]. Our evaluation shows that they are very likely incorrect. Since we do not know what information these figures are based on, we cannot pinpoint the cause. However, we assume that this count did not differentiate between the number of summonses and the number of individuals heard. It may also be that these figures do not differentiate between summonses served to individuals who were actually heard and to individuals whose summonses were canceled or who failed to appear. This is also indicated by the response of the Oberlandesgericht press office to our inquiry. According to court speaker Florian Gliwitzky, the “unverified” list of the press office mentioned a total of 814 summoned [sic!] witnesses and experts. Gliwitzky points out that the press office does not know, whether these witnesses and experts were actually heard by the senate; it is possible that individual witnesses were dismissed without being heard and then summoned again at a later time.
 There some are difficulties drawing the lines. Only the concrete initiators of summonses are counted. So if a witness was summoned upon request of the defense or the accessory prosecution, and if this summons made it necessary for the court to summon more witnesses, these additional summonses are counted as being initiated by the court.
 Of course this count of individuals could also be divided into smaller or broader groups. The “Criminal police” group could be divided into those officers who investigated the crimes committed by (but then not attributed to) the NSU or who investigated against the core trio before they went underground in 1998, and into those who investigated against the NSU after Zschäpe’s self-incrimination in 2011. Even with smaller categories, lines are difficult to draw. Most individuals in the “Supporters” group also fit into the “Nazis” group. Beate Zschäpe’s cousin, Stefan Apel, could be counted both as a former or active member of the Nazi scene or as a family member of the defendants (where we put him). We based our classification on which role the person primarily played in the trial. Signed and unsigned informers are classified as either “Nazis” or “Supporters”.
 The accusation of protracting the trial is just one in a number of media attacks against the lawyers of the accessory prosecution. For example, current Die Welt court reporter Gisela Friedrichsen wrote – while still working for Der Spiegel – of an “all-powerful forum of victims and their lawyers.” However, most voices accusing the accessory prosecution of protracting the trial have fallen silent by now.
 Individually: Tino Brandt (Verfassungsschutz Thuringia), Kai Dalek (Verfassungsschutz Bavaria), Marcel Degner (Verfassungsschutz Thuringia), Benjamin Gärtner (Verfassungsschutz Hesse), Andreas Rachhausen (Verfassungsschutz Thuringia), Thomas Starke (Criminal Police Office (LKA) Berlin), Carsten Szczepanski (Verfassungsschutz Brandenburg), Juliane Walther (Verfassungsschutz Thuringia).
At least for Kai Dalek, it remains unclear, if he is an informer in the classical sense: “He is said to have spied on the radical left for the local Verfassungsschutz office in West Berlin. The Berlin office then duly transferred its collaborator to the colleagues in Bavaria. Due to his ‘nationalist views,’ they preferred to deploy him in the right-wing scene. So Dalek is probably not a classic informer (V-Mann) but a sort of undercover agent for the Verfassungsschutz.” (“V-Mann-Porträt über Kai Dalek” [DE] by Robert Andreasch).
 It must be observed that individual witnesses and experts were counted repeatedly, if they were heard on different crimes. If they were heard multiple times on the same crime, they were just counted once.
 This figure is only an approximation. Counted strictly according to the Code of Criminal Procedure, much more applications/suggestions were made, since one individual application text often contains multiple applications or suggestions for a particular topic. For example, “alternative” applications are often made, in case the main application should be rejected. Or it is requested to summon a witness to testify on the contents of a particular document while also requesting the reading out of this document or the retrieval of a record relevant for the examination. Since counting all applications in this strict sense is difficult, each requested witness was treated as one application, but ancillary applications for retrieving or reading out documents were usually not counted extra.
 This figure might not be precise, since some motions for judicial disqualification might have been submitted outside of the main hearings. It must also be observed that only individual motions were counted, but not, for example, if another defendant joined the motion. If joint motions are counted, we arrive at 31 motions for judicial disqualification.
(Update 21.02.2018: We corrected the number of families of victims, whose members were heard as accessory plaintiffs: 4 instead of 5.)