United States v. Butts, 514 F. Supp. 1225 (M.D. Fla. 1981)

Lawrence Gentile, III, Asst. U. S. Atty., Jacksonville, Fla., for plaintiff.

Archibald J. Thomas, III, Asst. Federal Public Defender, Jacksonville, Fla., for defendants.

SUSAN H. BLACK, District Judge.

This cause is before the Court on defendant HARRY PICOT BUTT's Second Motion to Dismiss, filed herein on April 15, 1981. The Court heard oral argument on May 7, 1981, following which the parties were given additional time to submit supplemental briefs. Both defendant and the Government did so.

Defendant BUTTS moves to dismiss the indictment against him in the above-styled case pursuant to 28 U.S.C. § 1867(a) (1976), on the ground of substantial failure to comply with the provisions of the Jury Selection and Service Act, 28 U.S.C. § 1861 et seq. (hereinafter "the Act"), in the selection of the grand jury which indicted him (hereinafter "the Ellis grand jury").

As grounds for his motion, defendant alleges four separate violations of the Act:

1. The Ellis grand jury was not selected at random, as required by the Act and the Plan of the United States District Court for the Middle District of Florida for the Random Selection of Grand and Petit Jurors (hereinafter "the plan");

2. The names of the grand jurors were not publicly drawn by the Clerk, as required by the Act and plan;

3. The foregoing two procedures reflect regularly employed practices, not single deviations from the requirements of the law; and

4. The combination of the foregoing constitutes a substantial violation of the Act necessitating dismissal of the indictment.

At oral argument, the Court heard the testimony of three of the deputy clerks of the United States District Court for the Middle District of Florida, the testimony of a psychologist as to the random nature of the procedures employed by the Clerk in selecting the Ellis grand jury, and testimony of the Operations Manager of the Informations Systems Division for the City of Jacksonville. No credibility problems were apparent in the testimony of any of the witnesses, and the Court's Findings of Fact reflect the contents of the testimony substantially as presented to the Court.

1. Mrs. Llewellyn Landreth testified about the procedures used to construct the juror pool from which the Ellis grand jury was chosen. Mrs. Landreth has supervised jury selection since April, 1967. She described the process as involving three stages. In the first stage, juror names are transferred from the list of all eligible voters in the Jacksonville and Ocala Divisions to a "master wheel." In the second stage, names are transferred from the master wheel to a "qualified wheel," which contains the names of persons qualified for jury service. In the third stage, names are chosen from the qualified wheel for duty on particular grand or petit jury panels. The master wheel and the qualified wheel are both contained in a computer.^[1]

2. In the first stage of the process, the computer picks the names of jurors from the voter lists for the master wheel by use of a "starting number" and an "increment number." The increment number is found by dividing the number of jurors needed for the master wheel (specified in the Plan in 1977 as at least 15,000 for the Jacksonville Division) into the total number of registered voters. The starting number falls between one and the increment number. The starting number indicates the name of the first juror to be chosen on the list. Each subsequent name is obtained by adding the increment number to the starting number. For example, if the starting number is 37 and the increment number is 40, then the first juror chosen is juror number 37, the second is juror number 77, etc. Once the starting number is chosen, jurors are chosen systematically from the voter list according to the increment number. As *1227 specified in the Plan, the starting number at this stage was chosen by lot. During the 1977 refilling of the master wheel from which the Ellis grand jury was chosen, Senior United States Circuit Judge Warren L. Jones chose the starting number by lot.

3. In transferring names from the voter lists to the master wheel, the computer could make the actual selection only for Duval and St. Johns Counties, which maintain computerized voter lists. For the other counties in the Divisions, the Clerks chose names manually from the voter lists, using the starting number/increment number procedure. The names chosen were then keypunched and fed into the computer containing the master wheel. Supervisors of elections in the various counties supply the voter lists to the Clerk. However, the lists arrive in different conditions: some are computerized (Duval and St. Johns Counties), some alphabetized, some with the names listed in haphazard fashion. Once the master wheel is prepared, the Clerk never sees a printed list of names but receives only a computer tape.

4. In the second stage of the process, names are transferred from the master wheel to the qualified wheel, which contains the names of jurors qualified for jury service and not subject to an excuse or exemption. To fill the qualified wheel, questionnaires were sent to 6000 of the 60,000 jurors in the master wheel. Again, a starting number and increment number were chosen, the increment number as before, but the starting number by Mrs. Landreth herself, who testified that she simply picked a number between one and the increment out of her head without thought or hesitation. Once the questionnaires were returned, the Clerk separated those apparently qualified from those claiming excuse or exemption. A district judge reviewed each claimed excuse and exemption. Mrs. Landreth testified that several starting numbers were picked during the transfer process from master wheel to qualified wheel, but that she had no access to either the voter list or the master wheel list when starting numbers were chosen. She further testified that she had never been advised of any impropriety in her method of picking qualified jurors' names, nor heard of any court decision regarding such methods, until the decision of the United States District Court for the Northern District of Georgia in U. S. v. Alexander, No. CR 79-09 N (N.D.Ga., March 17, 1981). She further testified that no one had ever sought public access to any aspect of the jury selection process before defendant BUTT's motion was filed.

5. In the third stage of the jury selection process, a district judge orders the Clerk to choose a grand or petit jury panel from those names in the qualified wheel. As before, a starting number and increment number are chosen, this time by the jury selection clerk. Mrs. Suzanne Koch, the present jury clerk, testified that the Clerk posts notice of the upcoming selection in the public portion of the Clerk's office, on a bulletin board plainly visible to anyone entering through the public entrance door. Notice is posted several days before the starting number is chosen. Mrs. Koch's desk, where the number is chosen, is in plain view from the counter separating the public portion of the Clerk's office from the area reserved for court personnel. Once a starting number is chosen, a second notice is posted, this time specifying a date on which the panel would be chosen by the computer and indicating that the Clerk would make arrangements for anyone wishing to view the actual computer "run."

6. Mrs. Koch testified that she picked the starting numbers in the third stage of jury selection either out of her head or by asking a co-worker to pick a number between one and the increment. Although the Clerk's office had an alphabetized printout of qualified jurors from the master wheel, Mrs. Koch testified that she never consulted the list before picking a starting number. Since the increment number used usually did not divide evenly into the number of qualified jurors, a "remainder group" was created, consisting of those jurors remaining in the wheel when the increment had been divided into the total a whole number of times. Those jurors were then dropped from the selection of that particular *1228 panel. Although they had no chance to be picked in that selection, since they fell outside any incremental segment of the wheel, those jurors' names were put back into the wheel for the next drawing. (Mrs. Landreth testified that, following the Alexander decision, the remainder group problem had been addressed by enlarging the increment slightly to include those jurors formerly excluded when the quotient of increment into total was not a whole number. She also said that all starting numbers were now picked by drawing a number out of a box, rather than out of someone's head.)

7. Mrs. Victoria Clements, a deputy clerk and in 1979 the jury clerk for the Jacksonville and Ocala Divisions, testified that she had picked the starting number for the Ellis grand jury. At the time she did so, she occupied Mrs. Koch's present desk. She recalled only that she had either picked a starting number out of her head or asked a co-worker for a number in the proper range, using each procedure about half the time. She testified that she had posted public notice several days in advance of the number selection and of the computer run in Atlanta in accordance with the Clerk's regular procedure as described by Mrs. Koch (see Findings of Fact 5, supra).

8. Dr. Linda Foley, an associate professor of psychology at the University of North Florida in Jacksonville, testified for the defense. Dr. Foley, although she is not a statistician and holds no degree in statistics or mathematics, has done graduate work in applied statistics and computer science and teaches research methods in social psychology, where statistics are used extensively. Dr. Foley testified as to the random nature of the methods used by the Clerk to select juries. She defined the term "random" as indicating that "every person in the population would have an equal opportunity to be chosen." She examined a list of all starting numbers for juries selected in the Clerk's office from 1977 to the present, and made the following observations: the number one was never chosen, despite 97 opportunities to do so; none of the starting numbers fell within the highest quartile of the increment range; eight of the numbers accounted for 56 of the 97 choices, or 58% of the starting numbers; and most of the numbers appeared to be clustered toward the middle of the range. Each of these factors indicated to her that the starting numbers had not been chosen at random. More importantly, she said, a number chosen out of a person's head can never be chosen at random, since everyone prefers certain numbers and disdains others, even though this mental process often occurs only subconsciously. Most people have a favorite number as well. If a person were to pick a number out of his own head part of the time and ask a friend for a number otherwise, Dr. Foley said, this would amount to two nonrandom patterns mixed together without curing the overall nonrandom effect. The nonrandom selection of a starting number "infects" the entire jury selection process with nonrandomness, said Dr. Foley, since the process fails to comport with the definition that each person in the sample has an equal opportunity to be chosen. Dr. Foley said that she had no doubt that the starting number selection process in the case of the Ellis grand jury was not random.

9. During the cross-examination of Dr. Foley, the Government asked her to compare various methods of choosing samples on a random basis.

The Government: ... I wanted to ask you if you recognize, as I have read from time to time, a difference between the concepts purely random and statistical random. Is there a difference in the two?

Dr. Foley: I'm not sure.

G: Well, do statisticians recognize that the manner of selection that is prescribed in the plan, the so-called increment number method, and say a by lot drawing of the starting number, is that purely random or statistically random? Is there any difference?

F: I'm not sure how you're using the terms. There are two ways of taking a random selection. The most random way would be to choose not only the starting *1229 number but each number after that randomly.

G: And how would that be done, because that's what I'm getting at. How would it be done?

F: Well, if you had a wheel with all of the numbers in it, you would choose the starting number, and then put the starting number back in, unless you're going to eliminate it, roll it again, choose another number until you have done all 50 or ever how [sic] many you're going to select.

G: Not just have the increment number plow through the wheel?

F: Right, but generally statisticians accept the way it's done here, which is a systematic sampling where the starting number is random, and then using the increment.

G: So right away we back off from the definition of your pure random selection to at least something that is agreeable to the statisticians, the manner of the increment number going through, right?

F: Both of them are random selections. There are two different possibilities. The first description is very seldom used because it is so time consuming. The second one, as long as the starting number is chosen randomly, can pretty well be assumed to be a random selection unless the names are in a particular order, such as socioeconomic status.

G: But it's not as pure as it would be the former way?

F: I don't know that pure is the word.

G: Well, not as perfect, not as random in the most perfect sense of the meaning, unless we plug the starting number back in every time?

F: Well, they're both ways two different ways of selecting a random sample.

G: Would you say as an expert that by putting the starting number in every time, that is a better way to do it if it wasn't so unwieldly?

F: You would have probably a better sample of the population.

10. In order to clarify this colloquy, the Court questioned the witness after both sides had rested.

The Court: Over an infinite period of time each number in a random selection would come out the same, the same number of times; is that right?

Dr. Foley: Yes.

C: Any time that you place a parameter shorter than infinitely on a selection process, then each number will not come out the same number of times; is that right?

F: Right, the likelihood is they will come out approximately the same, and the more chances that you take

C: The closer you get to infinity, the more likely the same number will come out the same number of times, so the shorter the parameters, the less likely?

F: Right.

C: Thank you.

11. Mr. William Calcagni, Operations Manager, Information Systems Division of the City of Jacksonville, testified about public access to the computers which select federal juries. He said that there existed no predetermined policy about public access to the computers. Rather, any member of the public could view the selection "run" if arrangements were made beforehand. Mr. Calcagni further said that individual computer jobs were not scheduled at any particular time on a given date, but were simply set to run at some time overnight. The computers operate around the clock. He testified that he had personally conducted many tours of the computer facilities, but that until now he had never received any requests from anyone to open the computer room to persons interested in the selection of federal jurors. The Court finds that the facility having the jury-selection computer was accessible to the public, and that the computer itself could be viewed freely on prior request.

1. Of defendant's allegations regarding the jury selection process that selection of a starting number was not public, not random, amounting to regular practices constituting *1230 a substantial violation of the Act the Court will address the public nature of the selection first.

2. A remedy is not provided for every violation of the Act or plan. Rather, "Congress, recognizing that there would undoubtedly be error in the jury selection process that should not result in the dismissal of an indictment, left room for harmless error by providing that dismissal should lie only when there was a substantial failure to comply with the Act." U. S. v. Evans, 526 F.2d 701, 705 (5th Cir.), cert. denied, 429 U.S. 818, 97 S. Ct. 62, 50 L. Ed. 2d 78 (1976) (emphasis in original).

3. The substantiality of a violation of the Act is determined by weighing the alleged violation against the goals of the statute. U. S. v. Maskeny, 609 F.2d 183, 191 (5th Cir.), cert. denied, 447 U.S. 921, 100 S. Ct. 3010, 65 L. Ed. 2d 1112 (1980); U. S. v. Smith, 588 F.2d 111, 115 n. 22 (5th Cir. 1979); U. S. v. Carter, 568 F.2d 453, 455 (5th Cir. 1978); U. S. v. Davis, 546 F.2d 583, 589 (5th Cir.), cert. denied, 431 U.S. 906, 97 S. Ct. 1701, 52 L. Ed. 2d 391 (1977).

4. The purpose of [the Act] is to provide improved judicial machinery for the selection, without discrimination, of Federal grand and petit juries. Its aim is to assure all litigants that potential jurors will be selected at random from a representative cross section of the community and that all qualified citizens will have the opportunity to be considered for jury service.

H.R.Rep.No.1076, 90th Cong., 2d Sess., reprinted in [1968] U.S.Code Cong. & Ad. News 1792. To that end, the Act "embodies two important general principles: (1) random selection of juror names from the voter lists of the district or division in which court is held; and (2) determination of juror disqualifications, excuses, exemptions and exclusions on the basis of objective criteria only." Id. at 1793. Judicial construction of the Act has been in accord with this statement of principles. "Otherwise technical violations of the statute constitute `substantial failure to comply' when they affect the random nature or objectivity of the selection process." U. S. v. Kennedy, 548 F.2d 608, 612 (5th Cir.), cert. denied, 434 U.S. 865, 98 S. Ct. 199, 54 L. Ed. 2d 140 (1977).

5. Thus, in assessing the substantiality of the alleged lack of public access to the juror selection process, the Court must determine whether and in what way such lack of access impaired either the random nature or objectivity of the selection process.

6. The Fifth Circuit has had occasion to address the public-access question. In U. S. v. Dalton, 465 F.2d 32, 34 (5th Cir.), cert. denied, 409 U.S. 1062, 93 S. Ct. 570, 34 L. Ed. 2d 515 (1972), the Court rejected a public-access challenge, saying, "The selection [of the grand jurors' names] took place in one of the rooms of the clerk's office with only court personnel in attendance. However, it was a room which was open to the public. This was not an invalid procedure." Later cases elaborated on the public-access requirement in the special context of computer-generated jury lists. The most extensive discussion to be found is in U. S. v. Davis, 546 F.2d at 589:

It will not suffice for defendant merely to show that the public-access provision was violated. The exclusive ground for challenging jury-selection procedures under the statute is "substantial failure to comply" with the Act. 28 U.S.C. § 1867(a), (e). This circuit has applied the statutory standard. See United States v. Evans, 5 Cir., 1976, 526 F.2d 701, cert. denied, 429 U.S. 818, 97 S. Ct. 62, 50 L. Ed. 2d 78. Accord, United States v. Armsbury, D.Or., 1976, 408 F. Supp. 1130; United States v. McDaniels, E.D.La., 1973, 370 F. Supp. 298, aff'd sub nom., United States v. Goff, 5 Cir., 509 F.2d 825, cert. denied, 423 U.S. 857, 96 S. Ct. 109, 46 L. Ed. 2d 83 (1975). As this court observed in Evans:

Congress, recognizing that there would undoubtedly be error in the jury selection process that should not result in the dismissal of an indictment, left *1231 room for harmless error by providing that dismissal should lie only when there was a substantial failure to comply with the Act.

. . . . .

Id. 526 F.2d at 705. See H.R.Rep.No. 1076, 90th Cong., 2d Sess. 15, reprinted in [1968] U.S.Code Cong. & Admin.News, pp. 1792, 1805.

"Not every failure to comply with the provisions of the Jury Selection Act will constitute a `substantial' failure." Armsbury, supra, 408 F. Supp. at 1143. A complaint must rise beyond the allegation of a mere "technical deviation" or even a number of them. See Evans, supra, 526 F.2d at 703.

Determining the substantial compliance question requires that the alleged violations of the Act be weighed against the goals of the statute. See, e. g., Evans, supra, 526 F.2d at 705-06; Armsbury, supra, 408 F. Supp. at 1142. The major policy of the Act is that juries shall be "selected at random from a fair cross section of the community." 28 U.S.C. § 1861. Discrimination is prohibited in the selection process. Id. § 1862. Toward these ends, voter lists are the primary source of jurors' names, and the Act aims to determine disqualifications, excuses, exemptions, and exclusions based solely on objective criteria. H.R. 1076, 90th Cong., 2d Sess. 4, reprinted in [1968] U.S.Code Cong. & Admin.News at p. 1793. Although the public-access problem constituted a technical violation of the statute, that difficulty in no way affected the random nature or objectivity of the selection process. There is no evidence that this shortcoming "operate[d] to frustrate the goals of the Act." Evans, supra, 526 F.2d at 706.

If we were left in the dark about the procedures employed behind closed doors, or if we had reason to believe that there were any improprieties which had escaped detection as a result of non-public drawings, the result might be different. In such a case, of course, the concern would be not with the procedures themselves but rather with their effect on the random nature and objectivity of the selection. But, in the absence of such circumstances or allegations, we hold that the lack of public access to the automated drawings did not amount to "substantial failure to comply" with the Act.

7. Seizing on the language, "if we had reason to believe that there were any improprieties which had escaped detection as a result of non-public drawings, the result might be different," defendant argues that, in the instant case, "lack of public access has indeed resulted in the impropriety of nonrandom selection having escaped detection." Def. Second Mot. to Dismiss at 4. The substantiality of the alleged nonrandom selection will be discussed infra. As to public access, the testimony established that public notice of starting number drawings and computer runs was regularly posted on a bulletin board plainly visible to anyone entering through the public entrance door several days before each drawing or computer run. This practice was in effect when the Ellis grand jury was chosen, and is in effect today. Defendant argues that no more than a few hundred people, at the most, could have passed through the Clerk's office during the time notice was posted. "Even if it be assumed that most of these people in fact saw the notice, this would still represent only a fraction of one per cent of the population of the Middle District of Florida." Def. Mot. at 8. Defendant points out that no notice was given in any county outside Duval, nor was the time or location of the computer selection in Atlanta specified in the notice.

8. "Publicly draw," in the context of jury selection, is a defined term in the Act.

(k) "publicly draw", as referred to in sections 1864 and 1866 of this chapter, shall mean a drawing which is conducted within the district after reasonable public notice and which is open to the public at large under the supervision of the clerk or jury commission, except that when a drawing is made by means of electronic data processing, "publicly draw" shall mean a drawing which is conducted at a *1232 data processing center located in or out of the district for which juror names are being drawn, and which is open to the public at large under such supervision of the clerk or jury commission as the Judicial Conference of the United States shall by regulation require; ...

28 U.S.C. § 1869(k) (Supp. III 1979). The Court now holds as a matter of law that the procedures employed in this case constituted sufficient public notice of the jury selection process. The Court in U. S. v. Davis remarked, "a complaint of lack of public access in the jury-selection process is not the sort of deviation from the statutory scheme that this court is prone to view as substantial noncompliance with the Act." 546 F.2d at 589, n.19. This is especially true in the case of a multi-purpose federal facility, like the building housing the United States District Court in Jacksonville.^[2] Persons having business before the Court can reasonably expect to find complete information about court proceedingsincluding jury selectiononly in the Clerk's office on the fifth floor, where notice was in fact posted. The fifth floor is also the place where the three active District Judges hold court. Therefore, it would not be logical to expect to find such information at any other place in the building. Although the United States District Court for the Northern District of Georgia did find that lack of public access constituted a substantial violation of the Act and local plan, its decision in U. S. v. Alexander is distinguishable in many respects. Judge Keady's detailed and scholarly discussion was careful to point out that the deputy clerk in charge of jury selection had persistently failed to post the required public notices over a significant period of time. When she did post notice, it was often only in her own private office behind her desk, and then no more than a few hours before the scheduled drawing. Moreover, failure to post public notice had been the subject of judicial attention in U. S. v. Davis. As the Alexander court put it,

Davis was decided in part on the assumption that the clerk's office had commenced to post proper notices of drawings and would continue to do so [citing 546 F.2d at 588 n.17] .... Davis' warning to the clerk's office went unheeded for, as we have found, inadequate notice continues even at this date. We resist the temptation merely to warn the clerk's office, as the Davis court did without resulting change.

Alexander at 53-54. No such circumstances exist in the instant case. No rebuke or reprimand was ever given to any personnel in the Clerk's office with regard to public notice, as Mrs. Landreth testified and the Court finds. The deputy clerk in charge of jury selection posted notice of starting number selection and computer runs in the public portion of the Clerk's office several days before the event. Actual selection of a starting number was made at a desk in plain view from the public section of the office. Had anyone inquired, arrangements would have been made to view the computer in operation as it selected jurors' names. Contrary to defendant's allegations, no one was "left in the dark about the procedures employed behind closed doors." 546 F.2d at 589. There has been no substantial violation of the Act or plan in this regard; no effect on either the random nature or objectivity of the juror selection process has been shown to be caused or masked by the alleged lack of public access.

9. This lack of untoward effect is more fully shown in considering defendant's second substantive charge, viz., that the grand jury was not selected at random. Defendant's attack focuses on the manner in which starting numbers were chosen by the deputy clerks. His expert, Dr. Foley, *1233 testified that nonrandom selection of a starting number infects the entire process from the start. Since no one can pick a random number out of his head, the argument goes, the entire subsequent selection of jurors flowing from such a nonrandom choice cannot be in compliance with the Act or local plan.

10. The term "starting number" nowhere appears in the Act. The plan mentions the term only in connection with selection of names for the master wheel from voter lists. Paragraph IV.D of the plan states, "The clerk shall ascertain the total number of registered voters for each division and divide that number by the number of names to be selected for the master jury wheel from that division .... Then he shall draw by lot a starting number ...." Mrs. Landreth testified, and the Court finds, that the Clerk complied with this provision of the plan when the master wheel was refilled in 1977, at which time Judge Warren L. Jones chose a starting number by lot. As to the drawing of names from the master wheel for the qualified wheel, however, and as to drawings from the qualified wheel for individual panels, paragraphs IX and X of the plan specify only that qualified wheels contain "the names of all persons drawn at random from the master jury wheels" and that the Clerk "publicly draw at random from the qualified jury wheel" names required for a particular panel. Therefore, the Clerk's failure to draw starting numbers "by lot" in choosing names for qualified wheels or particular panels does not in and of itself constitute a substantial violation of the Act, since the plan requires by lot drawings only when filling the master wheel.

11. Nor does the Clerk's failure to maintain an actual box or wheel, from which names may be drawn, constitute a violation of the Act or plan. The Fifth Amendment to the plan, approved by the Fifth Circuit Reviewing Panel on February 23, 1977, amends paragraph IV.E of the plan as follows:

The random selection of names of prospective jurors to serve on grand and petit juries from the voter registration lists of the counties comprising each division may be made by the Clerk either manually or through the use of a properly programmed electronic data processing system or device, or through a combination of manual and computer methods.

This comports with the statutory definition of "jury wheel," said to "include any device or system similar in purpose or function, such as a properly programmed electronic data processing system or device." 28 U.S.C. § 1869(g) (1976). It should be noted that the plan for the Northern District of Georgia prescribes that the starting number be drawn "by lot from a drum or box containing consecutively numbered cards covering the same range of numbers as the `quotient.'" Alexander at 41, 49. The Alexander court might well react sharply to its clerk's blatant disregard of an explicit directive as to the manner of selecting a starting number. The plan for the Middle District of Florida contains such a directive only in connection with selection of the starting number for the master wheela directive which was observed in choosing the Ellis grand jury.

12. Nor does the Clerk's use of an arguably nonrandom method of selecting a starting number necessarily taint the entire process, as defendant contends. Dr. Foley's testimony was to the effect that a starting number chosen nonrandomly renders the entire process nonrandom. In this regard it is interesting to contrast the testimony before Judge Keady in Alexander, where the Government's statistical expert testified that it is the increment number, and not the starting number, which is central to random selection of jurors' names:

Dr. Wang was of the firm opinion that despite these non-random starting numbers, over the passage of time the likelihood that a particular juror would be chosen was approximately equal to that of any other person in the qualified wheel. He characterized Turner's methods as nonrandom processes that nevertheless culminated in random results since Dr. Wang's analysis showed that *1234 "over a period of time," the actual starting number selected becomes irrelevant to the probability of any particular prospective juror being picked. He based his conclusion on the observation that increments are different for each selection because the qualified wheel constantly dwindles; because petit and grand juries requiring different panel sizes are drawn from the same list and therefore alter the relative order of jurors remaining in the wheel; and because the "set over" of unused petit jurors to another petit jury contributes to differing increment numbers. He therefore was of the opinion that the increment number and not the starting number is the key to randomness. This expert concluded that in view of these variables, the starting number over a period of time has little, if any, effect on the chance of any person in the qualified wheel being chosen as a venireman.

Alexander at 43-44 (emphasis supplied). Although Judge Keady "found as a fact that the jury clerk's starting number choices were not random," and that therefore "the selection process did not result in random panel choices for individual grand juries," id. at 44-45, the Court takes issue with any conclusion that such practice standing alone constitutes a substantial violation of the Act.

13. The legislative history of the Act reveals an intent by Congress to allow something short of randomness, as statisticians use the term, to suffice for the "random selection" contemplated by the Act. Judge Keady recognized as much in setting out, in a long footnote, an excerpt from the Senate Judiciary Committee report on the bill:

In the interest of theoretical accuracy, some qualifications should be added. Although the bill declares as its policy that litigants in the Federal courts "have the right to grand and petit juries selected at random," the selection system under the bill would not be entirely random. At various points in the process, candidates for jury service may be eliminated if they fall short of the requirements for service specifically enumerated in the bill. Moreover, peremptory challenges and challenges for cause are retained. All of these instances are, strictly speaking, exceptions to a purely random selection system. But the bill otherwise firmly establishes the principle of randomness by requiring the use of random techniques for arriving at the names of those who shall then be subject to the specified bases for elimination.

It is also true that, to the extent that the bill does provide for random selection, it does not insist upon randomness in the sense in which that term might be understood by statisticians. Many districts may seek the aid of statisticians in developing systems of selection that do meet the standards of that profession, and they are encouraged to do so. No doubt such systems enhance the likelihood of attaining the cross sectional goal of the bill. But for reasons of administrative feasibility your committee did not deem it necessary to require the use of random selection in the statistical sense. It is sufficient for the purpose of this legislation if the plan adopts some system of selection that affords no room for impermissible discrimination against individuals or groups.

Thus, for example, the plan may specify selection of every 70th name from the voter list, or every name at the bottom of a page in the list, or all the names on the list, even though in certain cases statisticians might not agree that truly random selection would be the result. Likewise, in drawing names from the mastet [sic] and qualified jury wheels some similar processes may be designated. Under this definition of randomness, the plan may also permit courts to continue to use "rotation methods" or "jury pools" in assigning persons to grand and petit juries, provided that these methods, too, are free from any taint of impermissible discrimination against groups or individuals ....

S.Rep.No.891, 90th Cong., 1st Sess. 16 n.9 (1967), quoted in Alexander at 52, n.38 (emphasis added). The House Judiciary Committee *1235 Report concurs. Although the Act sets forth basic criteria which each local jury plan must meet, "the plan approach is designed to provide a significant measure of flexibility so that localities may adjust the administration of jury selection to their particular needs." H.R.Rep.No.1076, 90th Cong., 2d Sess., reprinted in [1968] U.S.Code Cong. & Ad.News 1799. Thus § 1863(b) (3) of the Act

requires each plan to specify a procedure for random selection from the voter lists. These procedures will, therefore, vary from district to district in accord with local needs. Voter lists are kept in forms suitable for processing by computer in some of the larger districts, but most districts probably will have to rely on some method of manual selection analogous to picking, say, every 37th name on the lists. But each procedure must ensure that the names so selected are placed in a master jury wheel and that each political subdivision of a district or division is substantially proportionally represented in the master wheel.

Id. at 1800. Absent a showing that jurors were excused or exempted from service on the basis of nonobjective, nonstatutory criteria, or that a cognizable group of jurors was excluded from the jury selection process, the Court holds that nonrandom selection of starting numbers, without more, does not constitute a substantial violation of the Act or plan.

14. It is important to remember that use of nonobjective, nonstatutory criteria was found in the selection procedures used by the Clerk of the Northern District of Georgia, with the result that several categories of otherwise eligible jurors were improperly disqualified, excused or exempted from jury service. Twenty pages of Judge Keady's Opinion were devoted to a discussion of the alleged abuses, so widespread and pervasive were the effects of the Clerk's misinterpretations of the Act. See Alexander 55-74. What transpired in Alexander is not the case here, however. Defendant introduced no evidence to show any discriminatory effect on any cognizable group of prospective jurors. Any "remainder group" created by use of whole-number increments constitutes no such cognizable group, as Judge Keady himself recognized ("We view this deviation from statistical randomness as minimal," Alexander at 43a, n.31.) The Clerk has now addressed the remainder group "problem" by enlarging the increment slightly to eliminate "remainder" jurors, although even under past practice they were put back into the qualified wheel and thus eligible for selection on subsequent panels.

15. The requirement that a defendant show the exclusion of some cognizable group from the jury selection process, although apparently supported by the legislative history of the Act, cannot be called settled law in this Circuit. Judge Keady cites U. S. v. Kennedy for the proposition that a "litigant need not show prejudice to establish a `substantial failure to comply' with the Act":

[W]hen a statutory violation directly affects the random nature of the selection process, there is no need to show that the violation tended to exclude some cognizable group from that process .... A departure from the statutory scheme that directly affects the random nature of selection establishes a substantial violation independently of the departure's consequences in a particular case.

548 F.2d at 612, cited in Alexander at 47. However, as the Government points out in its Supplemental Response of May 13, 1981, the Fifth Circuit has declared otherwise: "Successful statutory challenge to a jury selection system hinges, in the first instance, on ... proof that the system is disadvantageous to a cognizable class, and second, proof that the disadvantage is occasioned by discrimination in the selection process." Foster v. Sparks, 506 F.2d 805, 819 (5th Cir. 1975). Judge Keady himself calls Kennedy's "pronouncements on violations of the Act vis-a-vis substantiality" "obiter dicta," Alexander at 48, and goes on to quote U. S. v. Smith, which distinguished Kennedy by saying that the "inherent random nature of the selection process is not significantly affected by such an occurrence" *1236 as took place under the facts of Smith. 588 F.2d at 115 and n.21, cited in Alexander at 48. The Court can only conclude that each case must rest on its own facts, but that absent a showing that some cognizable group was excluded from the jury selection process, no substantial violation will lie. Cf. U. S. v. Evans, 526 F.2d at 706:

The overwhelming majority of the errors alleged to have been committed by clerks in the screening process appear to have resulted not in the discriminatory exclusion from jury service against which the Act was intended to safeguard, but rather in the inclusion of persons who perhaps should have been relieved of jury duty. There may have been isolated examples of mistaken exclusion, but these alone do not afford a basis for challenging the grand jury.

(Italics in original) (Middle District of Florida plan).

16. Finally, it should be noted that even a selection process satisfying a statistician's definition of randomness may not result in a completely random outcome. This is borne out by Dr. Foley's testimony under questioning from both counsel for the Government and the Court, where she admitted that parameters shorter than infinity, placed on the frequency with which all names are chosen, may cause a disparity in the opportunities which particular names have in being selectedthus potentially violating her own definition of randomness, which demands strict equality of opportunity in the chance of selection of any juror's name. Likewise, although statisticians accept the starting number/increment number system as random, Dr. Foley admitted that a system in which the wheel was "rolled" each time a name was chosen would produce a better sample of the population. This points up an ambiguity inherent in the Act's use of the term "random": must the selection process only begin randomly, through use of randomly chosen starting numbers, or must the effect throughout comport with the most random method possible? The legislative history explicitly admits that some compromise in methodology is acceptable, so long as the system used "affords no room for impermissible discrimination against individuals or groups." S.Rep.No.891, cited in Alexander at 53 n.38. On the basis of the evidence adduced, the Court finds that neither the alleged lack of public notice nor the allegedly nonrandom selection of starting numbers caused the "inherent random nature of the selection process" to be "significantly affected," U. S. v. Smith, 588 F.2d at 115. Accordingly, no substantial violation of the Act or plan has been shown.

It is, upon consideration

ORDERED that defendant BUTT's Second Motion to Dismiss, filed herein on April 15, 1981, is denied.

[1] In 1977, when the master wheel for the Ellis grand jury was constructed, the computer was located in Atlanta, Georgia. Since October, 1980, the computer has been located in the City Hall, Jacksonville, Florida.

[2] Other federal agencies having bureaus or offices in the United States Post Office Building at 311 West Monroe Street include the United States Postal Service, United States Public Health Service, Veterans' Administration, Immigration and Naturalization Service, Border Patrol, United States Attorney, Federal Public Defender, United States Probation Office, United States Tax Court, United States Bankruptcy Court and United States Court of Appeals for the Fifth Circuit.