MEMORANDUM AND INTERIM ORDER ON DEFENDANTS’ MOTIONS TO DISMISS INDICTMENTS
The indictments at issue in the instant cases were handed down in March, April, May, and August of 1981. The defendants seek dismissal of the indictments on the grounds that Essex County grand juries, including the grand juries which indicted them, have been selected in a way which has resulted in the systematic underrepresentation of the minority groups to which they belong over a substantial period of time. The defendants argue that, as a result of this underrepresentation, the indictments violate rights secured by the equal protection clause of the Fourteenth Amendment to the United States Constitution and by Article XH of the Massachusetts Declaration of Rights.
In order to prove an equal protection claim, the defendants must meet their burden of establishing: (1) the group to which each belongs has been disproportionately underrepresented over a significant period of time in the grand jury pool; and (2) that the juror selection procedure is “susceptible of abuse or is not racially neutral.” Commonwealth v. Bastarache, Mass. Adv. Sh. (1980) 2465, 2474. The defendant must show that he is a member of the group which has been underrepresented. Id.; Castaneda v. Partida, 430 U.S. 482, 494 (1977). The defendants have established that 8 are black and 14 are Hispanic. See the table attached hereto. Thus, all of the defendants belong to “a recognizable, distinct class, 'singled out for different treatment under the laws, as written or as applied,” within the meaning of Castaneda v. Partida, supra at 494.
After the defendants have met their initial burden of proving underrepresentation, at least as to the federal equal protection claim, the burden then shifts to the Commonwealth to prove that this underrepresentation did not occur because of intentional discrimination. Castaneda, supra at 494. See also, Washington v. Davis, 426 U.S. 228, 239 (1976); Village of Arlington Heights v. Metropolitan Housing Development, 429 U.S. 252, 265 (1977). *599Under the State constitution even unintentional discrimination raises a constitutional question. Commonwealth v. Bastarache, supra at 2479.2 The coürt need not be concerned about these differences at this stage, however, until the defendants have met their initial burden of proof.
The court has held two sets of hearings on December 28, 29 and 31, 1981 and May 14 and 21, 1982. After hearing evidence from three statistical experts, the court now concludes that the defendants have not met their burden of proving underrepresentation of black grand jurors. Due to difficulty with data collection, however, the court will hold further hearings on June 21, 1982 to determine whether the defendants can prove Hispanics have been underrepresented on Essex County grand juries. At that time the Commonwealth may introduce further evidence to show that the methods used to select grand jurors were not “susceptible of abuse” or “racially neutral.”
I. The Sample
On motion of the defendants, a questionnaire was mailed to 328 individuals who had been summoned for grand jury duty in Essex County between 1976 and 1981. Though substantial numbers of those individuals actually served on grand juries, the record does not reflect how many of the total actually served or were summoned but failed to serve. Nonetheless, the parties agree that those 328 individuals represent the best available sample of the Essex County grand jury pool for the years in question.
The questionnaire contained the following two questions: first, asking the respondent’s race, and second, seeking his ethnic identification. On February 4, 1982, the parties stipulated that 331 questionnaires had been mailed; 23 were found undeliverable; and 260 were returned. Of the returned questionnaires, 155 were signed and answered; 105 of the questionnaires were answered and returned but left unsigned. Of the signed responses, none indicated that the respondent was either black or Hispanic. Of the 105 unsigned responses, one checked “Black,” and one checked both “Black” and “Hispanic.” The parties have not stipulated that the returns from the questionnaire accurately reflected the composition of the grand jury pool as a whole.
The Commonwealth argues that even if the questionnaire revealed zero black and Hisjpanic grand jurors included in the pool of 328, the black and Hispanic population of Essex County was so small that any conclusion of underrepresentation would not be statistically significant. To conserve resources, the parties filed a joint motion (styled a “Stipulation”) to ask that the court initially determine whether the defendants could sustain their burden of showing underrepresentation even if no blacks or Hispanics had appeared in the 328-member jury pool.
The court held hearings on that narrow issue on May 14 and 21, 1982, and the following experts testified: the defendants’ expert, Dr. Thomas J. Marx; the Commonwealth’s expert, Jeffrey W. Eiseman; and a statistical expert especially appointed by the court pursuant to Mass. R. Civ. P. 41, Dr. Michael Malik.
*600Defendants filed an affidavit of Dr. Marx on April 9, 1982. Commonwealth filed affidavits of Mr. Eiseman and Charles F. Grimes, Assistant District Attorney. The following exhibits have been received:
1. 1980 Census Advance Report, with definitions.
2. Extracts from Exhibit 1 for Blacks and Hispanics.
3. Distribution of Essex County Blacks and Hispanics by percentage.
4. Breakdown of Hispanic and Non-Hispanic.
5. Data Chart prepared from census document “summary tape file 1A (STF).”
6. Affidavit of Marx, referred to above.
6A. Addendum to the foregoing.
6B. Marx vita.
7. Eiseman affidavit, referred to above.,
7A. Eiseman revision of section 5.5.3.
7B. Eiseman revision of section 3.1.
8. Sketch by Eiseman.
Final Figures
The final figures were as follows by group, numbers and % of Total:
Blacks, 4045, 0.87%. '
Hispanics, 9225, 1.98%.
Total: 464,477
Grand Jurors summoned and/or served: 328
At-the May 14 and 21, 1982 hearings, the defendants offered some evidence not only on the statistically expected number of blacks and Hispanics in the grand jury pool, but also on the expected number of grand jurors from the “combined minority’ ’ of blacks and Hispanics taken together. The defendants have made no showing, however, that blacks and Hispanics taken together constitute a “recognizable, distinct class” for the purposes of either the Federal or State Constitutions. See, Commonwealth v. Peters, 372 Mass. 319, 322 (1977). Nor do the members of one minority group have standing to complain of underrepresentation of the members of other groups on grand juries. Though any defendant may complain that the petit jury that tries him is not drawn from a “fan cross-section of the community in that some group of which the defendant is not a member is under-represented, the “fair cross-section” requirement has generally been limited to petit juries. Taylor v. Louisiana, 419 U.S. 522 (1975); Duren v. Missouri, 439 U.S. 357 (1979); Commonwealth v. Bastarache, supra at 2467 n.1, 2474-2475. But see, Peters v. Kiff, 407 U.S. 493, 504 (1972), cited in Commonwealth v. Soares, 377 Mass. 461, 480-481 (1979) (noted for the point that a defendant hasv standing to challenge conviction on the grounds that blacks were arbitrarily excluded from both grand and petit juries) (plurality opinion).
The principal issues addressed at the hearing were: (1) whether Essex County’s population should be considered as a whole or sub-divided by communities for statistical and sampling purposes; (2) whether the expected number of Hispanic grand jurors should be reduced to reflect the possibility that disproportionate numbers of Hispanics might be aliens; and (3) what standard of statistical significance should be applied to determine whether significant underrepresentation has been shown.
For reasons that are set forth in more detail below, the court finds that the black defendants cannot sustain their initial burden; even if no blacks were included in the available sample they would not be underrepresented in the grand jury pool. The defendants who claim blacks are underrepresented cannot sustain their burden even if the unresolved factual questions were resolved in their favor; their motions to dismiss will therefore be denied. A finding that no Hispanics were included in the pool may be sufficient to establish significant underrepresentation, however, and thus further fact-finding is required in those cases. Several preliminaries must be addressed at the outset in order to properly analyze the data.
II. Statistical Test
The court must initially determine the proper statistical test to apply to the data collected by the parties. The Supreme *601Judicial Court has not yet adopted any particular method of analyzing the statistics. See, Commonwealth v. Bastarache, supra at 2481-2482. For the reasons below, the court holds that the so-called “.05 level of significance” test will be applied to the instant case.
Authorities differ on the standard of statistical significance to employ in cases such as the present one. In an of ten-cited footnote, the Supreme Court concluded that “as a general rule for such large samples, if the difference between the expected value and the observed number is greater than two or three standard deviations, then the hypothesis that the jury drawing was random would be suspect to a social scientist.” Castaneda v. Partida, supra at 496-497 n.17. This is not conclusive, however, because the Court did not state that the “two or three standard deviations” threshold is binding in future jury discrimination cases.
The Court refers in that footnote to other authorities' which advocate somewhat different threshold values of statistical significance but embraced the “statistical decision theory” using the .05 level of significance as “the value most commonly used by statisticians.” See Finkelstein, The Application of Statistical Decision Theory to the Jury Discrimination Cases, 80 Harv. L. Rev. 338, 359, (1966), cited in Castaneda, supra at 496-497 n.17. See also, Villafane v. Manson, 504 F. Supp. 78, 85-87 n.11, 14 (D. Conn. 1980) and authorities cited.
As a substitute for a statistical measure of underrepresentation, some courts have considered the arithmetic disparity between the percentage of minority group members in the population as a whole compared to the percentage in the grand jury pool. See e.g., United States ex rel. Barksdale v. Blackburn, 639 F.2d 1115 (1981) (en banc). See also, Turner v. Fouche, 396 U.S. 346, 359 (1970) (23 percent disparity prima facie evidence of discrimination); Carter v. Jury Commission, 396 U.S. 320, 327-328 (1970) (33 percent disparity prima facie evidence of discrimination); Hernandez v. Texas, 347 U.S. 475, 480-481 (1954) (14 percent disparity prima facie evidence of discrimination); United States v. Butler, 611 F.2d 1066 (5th Cir. 1980) (8.69 percent disparity between “whites” and “non-whites” acceptable). Many of these cases were decided in the wake of Swain v. Alabama, 380 U.S. 202 ( 1965), which found acceptable a (disparity of greater than ten percent, but before Castaneda, supra, which relied on the more precise method of statistical decision theory.
Recent decisions of this state’s highest court make it clear that, for State constitutional purposes, any departures from random selection which disfavor racial or ethnic groups require the closest scrutiny. Commonwealth v. Bastarache, supra at 2481-2482; Commonwealth v. Rodriques, 364 Mass. 87, 92 (1973). At least for State constitutional purposes, therefore, uncertainty about the appropriate statistical threshold is to be resolved in favor of the more sensitive test of discriminatory effect.
. Thedefendants argue.that a prima facie case of substantial underrepresentation is made out if the odds are not better than one in twenty that the observed Underrepresentation of blacks or Hispanics could have arisen by chance, the so-called “.05 standard.” The Commonwealth has suggested either an arithmetic measure of disparity such as that employed in Swain v. Alabama, supra, or a two standard deviation threshold. In order to utilize the more stringent test, the court finds the .05 level of significance standard is the more appropriate standard for testing the sufficiency of the defendants’ prima fade case.
After the December hearings, the court severed for initial determination the following two issues: (1) whether the group to which the defendants belong has been substantially underrepresented on Essex County grand juries during the five-year period from 1976 to 1981, and (2) by what process were Essex County grand jurors selected during that period. These issues are now examined In turn.
III. Underrepresentation
The defendants must prove that the *602grand jury pool substantially underrepresented the Hispanic or black populations in Essex County.
A. Data as to Black Defendants
Both the defendants and the Commonwealth relied on 1980 United States Census Figures for Essex County whose accuracy was stipulated. The census figures showed that this total adult population of Essex County in 1980 was 165,477. The total adult black population was 4043. The black adult population was thus approximately .87% of the entire adult population of the county. Dr. Marx, the defendants’ expert, testified that the expected number of blacks in a random sample of 328 individuals drawn from the population in question would be slightly less than three (about 2.85). On cross-examination, Dr. Marx admitted that the probability of obtaining blacks in a random sample of 328 would be about .054. He also testified that, if the court adopted a strict .05 standard then the complete absence of black grand jurors from the sample would be insufficient to establish statistically significant underrepresentation. No evidence contrary to this conclusion was presented. On the basis of this testimony, the court finds that, even if the available sample contained no black grand jurors, the defendants would be unable to make out a prima fade case of discrimination against blacks. ■"
B. Data as to Hispanic Defendants
Again, the census figures showed that the total adult population of Essex County in 1980 was 165,477. The total adult Hispanic population, including both Puerto Ricans and non-Puerto Ricans, was 9225. The adult Hispanic population was approximately 1.98% of the total. The expected number of Hispanics would be, according to Dr. Marx’s testimony, somewhat less than 7 (about 6.5). Dr. Marx testified, and the court finds, that for the population in question, a minimum number of three Hispanics would be consistent with any random drawing at the .05 level of significance. In other words, if the court found 0, 1, or 2 Hispanics out of the 328 sample of grand jurors, the underrepresentation of Hispanics would not be due to chance.
The probability that there would be no Hispanics at all in a random sample of 328 grand jurors drawn from the relevant population would be one in 198. That is, if- there were no Hispanics at all in the sample, one would be justified in concluding at the relatively stringent .05 level of significance that there was systematic underrepresentation. Therefore, clearly, at least at the .05 level of significance, the one Hispanic in the 328-member sample (assuming the sample is an accurate representation of the grand jury pool) would constitute significant underrepresentation.
IV. Process of Juror Selection
A. Community-by-Community Standard
The Commonwealth attacks the validity of the sample on the basis that the grand jury selection process would be judged by a community-by-community standard when each city or town selects the jurors. The court holds that the defendants’ complaint will be considered on the basis of underrepresentation on the grand jury of the whole county.
The Commonwealth argues that selection from the county as a whole distorts reality in two respects. First, it failed to make an adjustment for that fraction of the Hispanic population which is ineligible for grand jury duty by reason of alienage or otherwise. Second, it ignores the “community-by-community” grand juror selection system which is apparently followed in Essex County.
The first of these arguments was put forward by the state defendant in Castaneda v. Partida, supra, but was rejected by the Supreme Court. Castaneda is clear that, once the defendant has shown underrepresentation by reference to the relevant minority group’s proportion in the general population, it is up to the state to rebut the inference of the discrimination “if it can” by reference to alienage, sound mind, moral character, criminal record, *603or other factors. Castaneda v. Partida, supra at 498-499. Moreover, in the present case, the Commonwealth failed to introduce evidence that the proportion of ineligible individuals was greater among the Hispanic population than among the population in general.3
The Commonwealth also argues that the defense’s calculation should have been adjusted to take account of the fact that a large fraction of the Hispanic population of Essex County lives in relatively few urban centers. The Commonwealth contends that it has been the practice in Essex County to draw only one or two grand jurors from a given city or town to sit on a given grand jury. The Commonwealth argues that this method of grand jury selection results in underrepresentation of residents of large cities and towns in general. The Commonwealth’s expert testified that, for a sample drawn not on a county-wide random basis, but on a community-by-community basis, the probability of obtaining 328 grand jurors none of whom were Hispanic would be about one in 36 rather than one in 198.
Data on the geographic distribution of grand jurors in the sample both parties agreed to adopt does not support the Commonwealth’s suggestion that grand jury members were selected according to a consistent allocation by community (see Exhibit #5).. Even if Essex County systematically discriminates only against dty-dwellers this fails to rebut the defendants’ claim of systematic underrepresentation, particularly if one effect of the system is to reduce the likelihood that Hispanics will be selected.
Finally, and perhaps most persuasively, the defendants were indicted by grand jurors from the county as a whole. The grand jury sits by county — not by city or town. Furthermore the enabling statute, G.L. c. 277, secs. 1, 3, speaks in terms of recruiting grand jurors from the county as a whole. None of the Commonwealth’s arguments warrants the selection of any geographic unit smaller than the entire county for the application of the relevant statistical test where the defendants’ complaint is of underrepresentation on the grand juries of Essex County, not of discrimination at the local level.
B. Key Man System
It is settled that the “key man” system of jury selection, under which unfettered discretion is vested in one individual or in a few individuals to select the members of the panel, is susceptible of abuse. Commonwealth v. Bastarache, supra at 2480. Castaneda v. Partida, supra at 497, The Commonwealth conceded at oral argument that at least two of the towns sharing in the selection of grand jurors in Essex County continue to use the “key man” system. Other towns, it is contended, use. other systems, and the Commonwealth maintains that it would be impossible to form an accurate picture of Essex County grand jury selection without, at a minimum, interviewing local officials from each of the 34 cities and towns in the county.
The Commonwealth’s admissions to date, however, are sufficient to support a finding that the Essex County grand jury system is susceptible of abuse or is not racially neutral. The Commonwealth admits that the selection system discriminates against residents of the large cities and towns and that it is in those cities and towns that the bulk of the Hispanic population lives. The Commonwealth also concedes that grand jurors are selected at the local level by a variety of methods not subject to central oversight or control, and among these methods is the “key man” system. This is enough to establish that the grand jury selection system in Essex County may be susceptible of abuse.
*604Order
For the reasons stated above, it is hereby ORDERED that (1) the motions to dismiss filed by the following defendants are DENIED: John Doe, aka, William Coles, Jr. (No. 2397); Homer Johnson (Nos. 2381-2382); Ronald Pittman (Nos. 2541-2542); Frank Tucker, aka Frank Pettiford (Nos. 2407-10, 2827); Dolphus Taggart (Nos. 2463-6); Dwight White (No. 2446); Levon Ferris (No. 2403); Willie Reed (Nos. 2455-7); Reginald Conley (Nos. 2401-2) and Kenneth Harris (Nos. 2383-7).
(2) A hearing is to be held on June 21, 1982 at which any further evidence which the parties wish to offer on the issues (i) whether Hispariics have been significantly underrepresented on Essex County grand juries between 1976 and 1981 and (ii) whether or not the Essex County grand jury system is susceptible of abuse. The court notes that if the defendants succeed in presenting their prima fade case, then the burden shifts to the Commonwealth to prove that the underrepresentation was unintentional. The court intends to try all these issues at the June 21 hearing.
By the Court,
Robert J. Hallisey J ustice of the Superior CourtClaim
JOSE APONTE, et al, Nos. 2371-2-3-4-5
2) Jorge L. Barbosa, 2420-21
3) Jose Cayas, aka Jose Colon Diaz, 2500
4) Jose M. Contreras, 2368-9-70
5) Felix DeJesus, 2379-80
6) John Doe aka William Coles, Jr., 2397
7) Felipe Diaz, 2376-7-8 and 2549-50-51
8) Moses Felix, 2448
9) Marshall Hall, 2395
10) Elijah Johnson, 2450
11) Homer Johnson, 2381-2382
12) Michael Knikker aka Nicholas Santiago, 2438
13) Jose Pichardo, 2439-40-41
14) Ronald Pittman, 2541-42
15) Otero Edgardo Suarez, 2435
16) Hiram Zaragosa, 2447
17) Frank Tucker aka Frank Pettiford, 2407-8-9-10 and 2827
18) Dolphus Taggart, 2463-4-5-6
19) Dwight White, 2446
20) Levon Farris, 2403
21) Julio R. Martinez, 2458-9
22) Willie Réed, 2455-6-7
23) Guilliermo Gonzales, 2363-4
24) Carlos Sanchez, 2553-4-5-6-7
25) Cesar Rios, 3207-8
26) Reginald Conley, 2401-2
27) Kelley Ferrer aka Camello Oterio, 2436-7
28) Kenneth Harris, 2383-4-5-6-7
FINDINGS OF FACT ON SEVERED ISSUE NO. 1 — RACIAL OR ETHNIC CLASS OF DEFENDANTS ON DEFENDANTS’ MOTION TO DISMISS BECAUSE OF COMPOSITION OF GRAND JURY
This severed issue came on for hearing 12/28 & 29/81.
Agreement filed 12/28/81
Nothing on Docket to indicate allowance to come back on motions. In Court on Record.
. The court notes that while under the State claim even unintentional discrimination may be prohibited the standards under both the state and federal constitution may be similar for two reasons. First, when jury lists are generated by a non-random process such as the “key man” system, as was the case with at least two cities in this case, the process is inherently suspect and therefore subject to special scrutiny. Bastarache, supra at 2470, 2480-2482. Second, the court has adopted the most stringent statistical test which may compensate to some degree for the differences in proving the state and federal claims. See, Villafane v. Manson, 504 F. Supp. 80-81 (D. Conn. 1980).
. See G.L. c. 234, 88 1, 1A (exemptions from jury service).