ORDER
The opinion and dissent filed January 21, 1994, slip op. 599, and appearing at 14 F.3d 1387 (9th Cir.1994), are withdrawn. A new opinion and dissent are filed in their place, and the petitions for rehearing and suggestions for rehearing en banc are dismissed as moot without prejudice.
OPINION
HARLINGTON WOOD, Jr., Senior Circuit Judge:
A federal grand jury indicted defendants Christopher Armstrong, Aaron Hampton, Freddie Mack, Shelton Martin, and Robert Rozelle for conspiring to distribute cocaine base in violation of 21 U.S.C. § 846 (1988). Some of the defendants also were indicted on substantive cocaine base charges under 21 U.S.C. § 841(a)(1) (1988), and using a firearm in connection with drug trafficking under 18 U.S.C. § 924(c) (1988 & Supp. Ill 1991). The defendants moved for discovery on whether the government selected the defendants for prosecution because of their race, and the district court granted the motion. After denying the government’s motion to reconsider, the district court dismissed the indictments as a sanction for failure to comply with the discovery order, but stayed the execution of the dismissals pending appeal by the government.
We have jurisdiction to hear the government’s appeal from the final judgment of the district court pursuant to 28 U.S.C. § 1291 (1988) and 18 U.S.C. § 3731 (1988). For the reasons stated, we reverse.
I. FACTUAL BACKGROUND
A task force composed of Inglewood Narcotics Division detectives and Bureau of Alcohol, Tobacco, and Firearms (ATF) agents used three confidential informants from February through April of 1992 to infiltrate a cocaine base1 distribution ring. On seven occasions from February 13, 1992, to April 6, 1992, the informants purchased cocaine base totalling approximately 124.3 grams from the defendants. The informants also reported the use of multiple firearms by the defendants during the sales.
On April 8,1992, task force police executed search warrants on the hotel room in which the informants made their purchases, as well as on residences belonging to some of the defendants. The officers arrested defendants Armstrong and Hampton in the hotel room, discovering 9.29 additional grams of cocaine base and a loaded gun. The officers subsequently arrested defendants Mack, Martin, and Rozelle pursuant to bench warrants the district court issued. Ultimately, the task force police seized multiple firearms and approximately 135 grams of cocaine base as a result of the investigation. All of the defendants are black.
The government sought indictments against all defendants in federal court. On April 21, a grand jury indicted all defendants [1433]*1433for conspiracy to distribute cocaine base under 21 U.S.C. § 846. The indictment also charged some defendants with substantive cocaine base violations of 21 U.S.C. § 841(a)(1), and usage of a firearm in connection with drug trafficMng in violation of 18 U.S.C. § 924(c). The federal statutes at issue provide for more stringent penalties than their California counterparts.2
On July 20, 1992, defendant Martin filed a Motion for Discovery and/or Dismissal of Indictment for Selective Prosecution, claiming that the government was prosecuting him because of his race. Defendants Armstrong, Mack, Hampton, and Rozelle all timely joined defendant Martin’s motion. The district court held a hearing on the motion on September 8, 1992.
At the hearing, the defendants offered as evidence of selective enforcement an affidavit from a paralegal employed by the Office of the Federal Public Defender. The affidavit, which included a statement and a chart, asserts that in the 24 cases closed by the Federal Public Defender’s Office in 1991 involving cocaine base violations of 21 U.S.C. § 841 and/or 21 U.S.C. § 846, the defendant in each case was black. The defendants for some reason did not offer an affidavit from the Federal Public Defender or any supervising attorney, or for that matter any other evidence at all, but instead relied solely on the affidavit from the paralegal employee. As a result, the government contended that the defendants failed to meet the showing required to compel discovery.
Nevertheless, on September 8, 1992, the district court disagreed with the government and granted the motion for discovery on the issue of selective prosecution. The district court ordered the government to: (1) provide a list of all cases from the prior three years in which the government charged both cocaine base offenses and firearms offenses; (2) identify the race of the defendants in those cases; (3) identify whether state, federal, or joint law enforcement authorities investigated each case; and (4) explain the criteria used by the U.S. Attorney’s Office for deciding whether to bring cocaine base cases federally.
On September 16, 1992, the government filed a motion for reconsideration of the discovery order. In support of its motion for reconsideration, the government submitted sworn declarations of a Special Agent of the Drug Enforcement Administration with 21 years experience, a Special Agent of the Bureau of Alcohol, Tobacco, and Firearms with three years experience at the ATF and another three years as a narcotics officer, a narcotics detective from the Inglewood Police Department with 10 years on the force and three years experience in the narcotics unit, and two experienced Assistant United States Attorneys stating that: (1) the Office of the Federal Public Defender represented at least five non-black cocaine base defendants during the relevant time period; (2) the government prosecuted many non-black cocaine base defendants during 1991, the period at issue in the report prepared by the paralegal employed by the Office of the Federal Public Defender; (3) the county district attorney’s offices prosecute many black cocaine base offenders; (4) the government based its decision to charge on the existence of federal firearms and narcotics violations that met the guidelines of the United States Attorney’s Office, the strength of the evidence, the deterrence value, the federal interest, the suspects’ criminal history, and other race-neutral criteria; and (5) socio-economic factors account for the prevalence of drugs in certain communities, as illustrated by black gangs in the south-central Los Angeles area predominantly controlling the supply of cocaine base.
In response to the government’s motion for reconsideration, the defendants offered [1434]*1434two additional declarations. The first, made by one of the defense attorneys, states that she had spoken with a halfway house intake coordinator who told her that in his experience in treating cocaine base addiction, the number of Caucasian and minority users and dealers is equal. The other declaration, made by another defense attorney, asserts that (1) he has represented only blacks in federal court on cocaine base charges; (2) he has never heard of non-blacks being prosecuted in federal court on cocaine base charges; and (3) in his conversations with unnamed state court judges, prosecutors, and defense attorneys, he has come to believe that the state prosecutes many non-black cocaine base offenders in state court. The defendants also submitted an article from the Los Angeles Times, which contends that blacks disproportionately commit cocaine base offenses. See Jim Newton, Harsher Crack Sentences Criticized as Racial Inequality, Los Angeles Times, Nov. 23, 1992, at Al, A20.
After a hearing on the motion for reconsideration, on December 29, 1992, the district court denied the motion. The government notified the district court on January 5, 1993 of its intention to challenge the discovery order and the denial of the reconsideration motion. As a sanction for failure to comply with the order, the district court dismissed the indictments of all defendants. The district court stayed the execution of the dismissals pending this appeal.
II. DISCUSSION
The parties did not discuss the relationship between two decisions of this court published within days of each other, United States v. Redondo-Lemos, 955 F.2d 1296 (9th Cir.1992), which was finalized on May 11, 1992, and United States v. Bourgeois, 964 F.2d 935 (9th Cir.1992), decided on May 19, 1992. Bourgeois and Redondo-Lemos, both thoughtful opinions, examined somewhat differently the process by which defendants can obtain discovery on a claim for selective prosecution.
In Redondo-Lemos, the defendant was caught transporting 695 pounds of marijuana into the United States from Mexico. The defendant pled guilty to a violation of 21 U.S.C. § 841(a)(1) in exchange for the government’s promise to recommend that the district court impose only the mandatory minimum sentence of five years imprisonment. The district court, without any motion by the defendant, held that the United States Attorney’s Office was selectively enforcing the drug laws against male defendants, and sentenced Redondo-Lemos to 18 months imprisonment instead of the 5-year mandatory minimum sentence.
On appeal, this court set forth the process for resolving a selective prosecution claim. First, Redondo-Lemos states that “[n]o one is entitled to call the prosecution to answer for a particular charging or plea bargaining decision without making a prima facie showing that wrongful discrimination is taking place.” Redondo-Lemos, 955 F.2d at 1302. Redondo-Lemos states that a prima facie showing can be satisfied in one of two ways: (1) if “the district court develops a suspicion of unconstitutional conduct on the basis of its own day-to-day observations,” or (2) if the defendant, rather than the district court, is the one to raise the claim of selective prosecution, “he must present enough evidence to demonstrate a reasonable inference of invidious discrimination.” Id. Only the suspicion of the district court, however, was at issue or applied in Redondo-Lemos.
Second, Redondo-Lemos states that once the prima facie ease is established, the Office of the United States Attorney must have an opportunity to rebut the prima facie case. At this stage, the showing that the United States Attorney must make “would not involve file information about specific cases, but [rather would] consist only of overall case statistics.” Id. Those overall statistics would be available to both the district court and defense attorneys. Id.
Third, if after the district court has seen the rebuttal evidence, the court finds discriminatory impact by a preponderance of the evidence, it then must determine if the prosecutor’s charging decision was based on a discriminatory motive. Id. That decision “may, in rare instances, extend to in camera examination of certain prosecution case files [1435]*1435and to limited discovery by opposing counsel.” Id. In this third stage Bedondo-Le-mos for the first time raises the possibility of discovery by the defense. If, once the district court examines discriminatory motive and listens to government rebuttal evidence, the court finds by preponderance, of the evidence intentional discrimination based on a suspect classification, it may fashion an appropriate remedy. Id.
Unlike Bedondo-Lemos, Bourgeois confines itself to the question of when a defendant may obtain discovery based on a claim of selective prosecution. At issue in Bourgeois was a 2-day series of firearms arrests known as “Operation Streetsweep.” Bourgeois, 964 F.2d at 986. Bourgeois, a felon, was caught in Operation Streetsweep and indicted for possession of a firearm in violation of 18 U.S.C. § 922(g)(1). Id. at 937. Bourgeois moved to dismiss the indictment based on selective prosecution, arguing that he was selected for prosecution because he is black. Id. Bourgeois also filed a discovery request for information regarding the sting operation in which he was arrested. Id. The district court denied the discovery request, and Bourgeois entered a conditional guilty plea. Id.
On appeal, this court examined the question of when defendants are entitled to discovery on selective prosecution claims. Bourgeois noted that to be ultimately successful on a selective prosecution claim, a defendant would have to prove “that others similarly situated have not been prosecuted and also that the prosecution is based on an impermissible motive.” United States v. Wayte, 710 F.2d 1385, 1387 (9th Cir.1983), quoted in Bourgeois, 964 F.2d at 938. Bourgeois then explained that it was an open question as to whether the standard for obtaining discovery was different from the standard for ultimate success on a selective prosecution claim, and to what degree. Bourgeois, 964 F.2d at 938.
The Bourgeois court decided that defendants must satisfy a high threshold to obtain discovery. Id. at 939. The court adopted a high threshold first because “courts are ill equipped to assess a prosecutor’s charging decisions,” and second because “court oversight of prosecutorial decisions could undermine effective law enforcement.” Id. The court sought to “discourage fishing expeditions, protect legitimate prosecutorial discretion, safeguard government investigative records, and yet still allow meritorious claims to proceed.” Id. at 940. The standard the court adopted to accomplish these ends was that “to obtain discovery on a selective prosecution claim, a defendant must present specific facts, not mere allegations, which establish a colorable basis for the existence of both discriminatory application of a law and discriminatory intent on the part of the government.” Id. at 939. Bourgeois noted that the “colorable basis” standard was meant to continue the trend of the past few decades, during which time only a handful of defendants obtained discovery on a selective prosecution claim. Id. at 940.
Thus, the question for us here is whether to apply the Bedondo-Lemos “suspicion” test, the Bedondo-Lemos “reasonable inference” test, or the Bourgeois “colorable basis” test. In our view, the record does not support the proposition that the district court based its discovery order on its day-to-day observations, and therefore the Bedondo-Le-mos “suspicion” test is inapplicable. That conclusion is grounded in the district court’s reasoning for granting the defendants’ request for discovery.
The district court, in explaining its decision to order discovery, stated: “The court sees it different than Bourgeois, in that in this case we do have something more than mere allegations.” The court thus relied on the reasoning of Bourgeois in analyzing the defendants’ evidence, not on its own day-to-day observations. Additionally, the district court noted that at issue is “a fairly general charge — one that we see regularly in this courthouse — and whether it is coincidental or not, that out of the group that the public defender [proffered] all of them happen to be of the same racial group.” (Emphasis added.) This further illustrates the district court’s reliance on the evidence, not its own suspicions. The fact that the district court specifically expressed no opinion as to whether the racial composition of defendants represented by the Federal Public Defender was [1436]*1436a coincidence indicates not suspicion, but rather a lack thereof.
The district court concluded that “the number [of black defendants shown by the public defender] is adequate that [it] would at least require the government to provide some explanation. The time period is such that would require some explanation. The charges are the same or similar, and the race is the same in each case.” Those comments immediately precede the district court granting the motion for discovery, and are the basis for doing so. The record demonstrates that the district court based its decision on the evidence presented by the defendants, not on its own day-to-day observations. As a result, the Redondo-Lemos “suspicion” test is inapplicable, and we must choose between the Redondo-Lemos “reasonable inference” test and the Bourgeois “colorable basis” test.
Although neither phrase is easily susceptible to further definition, we believe that it would be more appropriate to apply the Bourgeois “colorable basis” test in this particular case than the Redondo-Lemos “reasonable inference” test. As previously noted, Redondo-Lemos did not apply the “reasonable inference” test, as the district court initiated the selective prosecution inquiry sua sponte. Additionally, in Redondo-Lemos the defendant was not seeking discovery. Redondo-Lemos, 955 F.2d at 1297. The court simply was attempting to explain the entire process for pursuing a selective prosecution claim, of which discovery is one part. See id. at 1297, 1302. We doubt that the court in Redondo-Lemos intended its discussion to be the final expression of law in discovery cases.
Bourgeois, on the other hand, addressed a situation similar to the one here. Bourgeois alleged that he was selected for prosecution based on his race, and filed a discovery request in an attempt to prove his claim. Bourgeois, 964 F.2d at 937. The court squarely confronted the issue of when a defendant is entitled to discovery, and devoted approximately three pages of analysis to that issue alone. See id. at 937-40. The court in Bourgeois analyzed a split in the circuits and supported its decision to apply the “colorable basis”-test with policy analysis. Id. When a district court bases its decision on evidence supplied by the parties rather than on its own day-to-day observations, Bourgeois is the law of this circuit regarding the test for determining whether to grant a defendant’s motion for discovery on a selective prosecution claim.3
Having decided that the Bourgeois color-able basis test governs here, we must now apply that test to the facts of this case. The defendants submitted three affidavits and a newspaper article in support of their motion for discovery. The first affidavit, the statement and chart by the paralegal that was the only evidence the defendants initially offered, was that of 24 cocaine base cases closed by the Office of the Federal Public Defender in 1991, all 24 involved black defendants.
Taken alone, that evidence does not establish “specific facts, not mere allegations, which establish a colorable basis for the existence of both discriminatory application of a law and discriminatory intent on the part of the government.” Bourgeois, 964 F.2d at 939. To demonstrate discriminatory application, a defendant must provide a colorable basis for believing that “others similarly situated have not been prosecuted.” Wayte, 710 F.2d at 1387, quoted in Bourgeois, 964 F.2d at 941. The first affidavit demonstrates only that others have been prosecuted, not that others similarly situated have not — obviously, a total lack of evidence cannot constitute a colorable basis for believing that discriminatory application of the law exists.
Requiring defendants to provide a color-able basis for believing that others similarly situated have not been prosecuted is a réa-sonable requirement. “Selective prosecution” implies that a selection has taken place. If a defendant is part of a protected class, that alone does not provide a colorable basis for believing that a selection has taken place; nor does evidence demonstrating that other members of the protected class were prosecuted provide a colorable basis for so believ[1437]*1437ing. Rather, a defendant must supply a colorable basis for believing that others similar to him except that they are not in his protected class were not prosecuted. Without a colorable basis to believe that others similarly situated were not prosecuted, the most reasonable conclusion is that the defendant was selected for prosecution because the government believed the defendant committed the offense; the fact that the defendant is a member of a protected class is coincidental.
The evidence that all 24 cocaine base cases closed in 1991 by the Federal Public Defender involved black defendants is insufficient to provide a colorable basis to believe selection took place. The statistic in question does not speak to whether the Federal Public Defender had other cases involving white defendants that did not close during that period. Additionally, there is no information on whether any of the defendants in the 24 cases were similarly situated to the defendants in this case. The government claims it selected the defendants for prosecution in this ease because they committed cocaine base offenses while possessing firearms. If none of the 24 eases handled by the public defender involved firearms, the statistic is of little meaning, particularly in light of the government’s explanation. As a final criticism, the statistic tells us nothing of the racial composition of defendants charged with drug trafficking who may have had sufficient resources to retain their own attorneys.
The defendants contend, however, that the statistic is meaningful in light of the newspaper article and two affidavits supplied in opposition to the government’s motion for reconsideration. The article, from the Los Angeles Times, states that white cocaine offenders are being punished less severely because whites predominantly violate powder cocaine laws, while blacks predominantly violate crack cocaine laws, .the latter of which carry longer sentences than the former. See Jim Newton, Harsher Crack Sentences Criticized as Racial Inequality, Los Angeles Times, Nov. 23, 1992, at Al, A20. We agree with the district court’s determination that nothing in the article supports the defendants’ position; indeed, the article arguable cuts the other way. Perhaps the defendants were alluding to an argument that the federal sentencing scheme violates equal protection because of this disparity, but because they did not actually make the argument, we express no opinion on that matter here. Any fault in this regard, if fault indeed exists, does not lie with the United States Attorney, for the United States Attorney plays no role in drafting the United States Sentencing Guidelines.
The two additional affidavits, although not harmful to the defendants’ position like the newspaper article arguably is, are' not helpful either. One of the affidavits, made by an' attorney for one of the defendants, states merely that an intake coordinator at a Pasadena halfway house had told the defense attorney that “it was his experience in dealing with the treatment of cocaine base addiction, that there are an equal number of Caucasian users and dealers to minority users and dealers.” This statement is: (1) hearsay — the defendants certainly could have gotten an affidavit from the intake coordinator himself; (2) unsubstantiated; (3) without any indication of sample size, duration, or methodology of any kind; (4) without an indication of whether the magnitude of the sales is similar to those of the defendants in this case; (5) based on “users and dealers” rather than dealers who possess firearms; and (6) based on users and dealers being treated for addiction, without any evidence that such a sample is representative of dealers as a whole. The district court erred, in placing any weight on such flimsy evidence.
The other affidavit, made by another attorney for the defense, states that “I talk to many state court judges, prosecutors, and defense attorneys ... [and] [t]here are many crack cocaine sales cases prosecuted in state court that do involve racial groups other than blacks. A major percentage of sales of crack cocaine cases are prosecuted against Latinos in the state courts.” This statement: (1) is hearsay — there are no affidavits from any of the unnamed purported sources of this information; (2) fails to indicate whether the magnitude of the sales is similar to those of the defendants in this ease; (3) does not include [1438]*1438any information about cocaine base offenses involving the possession of firearms; and (4) gives no statistics, but rather merely reports the defense attorney’s personal conclusions that there are “many” eases and that there is a “major” percentage. These are hardly the “specific facts, not mere allegations, which establish a colorable basis for the existence of ... discriminatory application of a law” envisioned by Bourgeois. Bourgeois, 964 F.2d at 939.
Given that the ineffective affidavits and arguably damaging article, even taken together, did not contradict the government’s explanation for its decision to prosecute, we must conclude that the district court abused its discretion in concluding that the defendants’ evidence provided a colorable basis to believe that discriminatory application of a law existed. Bourgeois, as previously explained, set a high threshold for obtaining discovery, and expressed very sound policy rationales for . doing so. See id. at 939-40. Had the district court viewed the defendants’ evidence in light of the reasoning of Bourgeois, it should have concluded that the evidence failed to establish a colorable basis for believing that the government was engaging in selective prosecution.
As a final matter, one could argue that Bourgeois should be distinguished based on the duration of the government operations involved in each case. In Bourgeois, a 2-day operation was at issue, id. at 936, and in the instant case the defendants proffered a 2-month “study” by the Office of the Federal Public Defender. That distinction, by the very language of Bourgeois, is unimportant in this case.
The court in Bourgeois found that Bourgeois had failed to establish a colorable basis that others similarly situated had not been prosecuted. Id. at 941. It did this because “Bourgeois [made] no attempt to show that, in any span of time, the government has declined to prosecute similarly situated, non-black felons illegally in possession of firearms.” Id. (emphasis added). Thus, the court explained that regardless of the duration of the operation in question, defendants attempting to obtain discovery on a claim of selective enforcement must provide a color-able basis for believing that similarly situated felons not in their protected class were passed over for prosecution. The evidence in this case was unsatisfactory for the very same reason, and the duration of the operations in each case therefore is irrelevant.
III. CONCLUSION
The district court erred in finding that the evidence the defendants presented in this case established a colorable basis for believing that the government engaged in selective prosecution. The defendants’ evidence, which likely would be inadmissible in other contexts, was tenuous at best, and in one case even arguably was counterproductive. If more than this quantum of evidence is not required, district courts too often and unnecessarily could become immersed in the workings of a coordinate branch of government, to the benefit of neither. Accepting the defendants’ weak evidence and disregarding the sworn declarations of experienced federal agents and two Assistant United States Attorneys that explained and refuted the defendants’ statistics was error, and for that reason the decision of the district court dismissing the indictments is
REVERSED.