LYNCH, Chief Judge.

Petitioner Everett Connolly seeks the grant of habeas corpus vacating his state court convictions from 2006 for drug distribution and trafficking on the grounds that the state proceedings violated his federal Confrontation Clause rights, as articulated in Melendez-Diaz v. Massachusetts, 557 U.S. 305, 129 S.Ct. 2527, 174 L.Ed.2d 314 (2009). The Massachusetts Supreme Judicial Court (SJC) had acknowledged that there was Melendez-Diaz error but found it harmless under a standard equivalent to the federal standard under Chapman v. California, 386 U.S. 18, 24, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967).

On federal habeas review, the district court denied the petition, reasoning that the state courts had already found that the error was harmless and that the petitioner could not show sufficient injury under the highly deferential standards announced by the Supreme Court in Brecht v. Abrahamson, 507 U.S. 619, 637, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993), and Fry v. Pliler, 551 U.S. 112, 119-20, 127 S.Ct. 2321, 168 L.Ed.2d 16 (2007). See Connolly v. Roden, No. 09-11987, 2013 WL 139702, at *3 (D.Mass. Jan. 11, 2013).

We affirm. We conclude that under the Brecht standard of review, which is even more deferential than the ordinary standard of review under the Antiterrorism and Effective Death Penalty Act ("AEDPA"), 28 U.S.C. § 2254(d)(1), petitioner has failed to show the "substantial and injurious effect or influence" required to set aside the SJC's affirmance of his conviction. Fry, 551 U.S. at 116, 127 S.Ct. 2321 (quoting Kotteakos v. United States, 328 U.S. 750, 776, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946)).

I.

The vast majority of the key facts of this case are undisputed. Massachusetts law enforcement officers began investigating Connolly for suspected cocaine distribution in 2003. The investigation included surveillance and observation by police officers, which found Connolly making numerous apparent drug sales using his minivan. Commonwealth v. Connolly, 454 Mass. 808, 913 N.E.2d 356, 361 (2009). On two separate occasions in August 2004, an undercover officer made controlled purchases of crack cocaine from Connolly. During the first purchase, the undercover officer obtained what was later measured to be 1.2 grams of crack cocaine for $200. In the second, the officer obtained 0.56 grams of crack cocaine for $100.

In light of the controlled purchases, the police obtained an arrest warrant for Connolly and a search warrant to search Connolly's minivan and its occupants. Id. at 362. The warrants were executed on September 9, 2004. During the search of Connolly's minivan, the police found a "large ball" of crack cocaine,1 "wrapped in electrical tape and later determined to weigh 124.31 grams," wedged under the dashboard. Id.

Connolly was indicted by a state grand jury in December 2004 on two counts of cocaine distribution and one count of cocaine trafficking.2 See Mass. Gen. Laws ch. 94C, §§ 32A(c) (distribution), 32E(b)(3) (trafficking). At trial in August 2006, the prosecution introduced several pieces of evidence to prove drug type and, as to the trafficking charge, quantity.3 With respect to the controlled purchases, Detective Lieutenant John Allen, who had 25 years of experience in narcotics, testified that he had personally field-tested the substances at the time of each of the two August 2004 purchases and concluded that they were cocaine; the undercover officer who made the purchases also testified that Connolly had said the substances were crack cocaine. Connolly, 913 N.E.2d at 375. With respect to the cocaine found during the minivan search, Sergeant John Milos, who had investigated hundreds of cocaine distribution cases over seventeen years of experience, testified that field tests showed the ball recovered from the van was cocaine. Id. Trooper James Bazzinotti, who had thirteen years of experience in conducting drug searches, further testified that the substance was cocaine, and that the ball looked "like a baked potato," was "bigger than a baseball," and was a "large ball, hard," and was "a very solid piece" of cocaine. Id. at 375-76. The jury also had direct physical access to the actual cocaine, which was sent to the jury room, although the ball was "more flaked, more crushed up" after the chemical analysis. Id. at 376.

The Melendez-Diaz problem in the case arose because the prosecution also introduced a certificate from the drug analysis laboratory explaining that the powder in the ball was cocaine and weighed 124.31 grams, without calling the analyst as a witness and so not making the analyst available for confrontation. At the time the certificate was admitted, the judge instructed the jury that under Massachusetts law, the prosecution was not required to offer live testimony from the analyst and could introduce the certificate alone without any adverse inferences being drawn due to the analyst's absence. Connolly objected to the admission of the certificate on Fourth Amendment grounds but not on Confrontation Clause grounds. He did not ask to cross-examine the analyst who prepared the lab certificate, nor did he object to the jury instruction on Confrontation Clause grounds. The closing jury instructions allowed the jury to consider the laboratory certificate but cautioned that "[i]t's for you to determine whether you credit it or not.... [Y]ou're not compelled to accept it."

Connolly's theory of the case was that the prosecution had not adequately proven that he was the drug possessor. He identified potential weaknesses in the testimony of the government's witnesses to argue that there was insufficient evidence that he was actually involved in the two controlled purchases or that he was aware that the ball of crack was in his van. Connolly made no challenge to the quantity of crack for purpose of the trafficking charge. In fact, in closing arguments, his attorney appeared to accept the quantity at face value and challenged only identity, arguing: "[t] he Commonwealth has to prove to you beyond a reasonable doubt that Mr. Connolly knew that those 124 grams w[ere] in his vehicle on the morning in question. I suggest to you there's no proof of it."

The jury convicted Connolly on all counts. Following the conviction, Connolly appealed to the Massachusetts Appeals Court. In January 2009, the SJC transferred the case to its own docket sua sponte for direct review. While the case was pending in the SJC, the U.S. Supreme Court decided Melendez-Diaz v. Massachusetts, which held that, absent a showing of unavailability, the introduction of a lab certificate without the opportunity to cross-examine the technician who created it violates a defendant's Sixth Amendment Confrontation Clause rights. 557 U.S. at 311, 129 S.Ct. 2527. Following Melendez-Diaz, Connolly presented the narrow argument that his conviction must be reversed because the trial judge had given a contemporaneous jury instruction upon the introduction of the certificates stating that the prosecution was not required to produce the lab analyst, in violation of his Confrontation Clause rights. He had not raised that argument at any earlier point in the litigation.

The SJC read Connolly's argument charitably in his favor by considering not only the instruction but also whether admission of the certificates themselves was error. See Connolly, 913 N.E.2d at 374-75. It decided that there was Melendez-Diaz error in admitting the certificates without presenting the analyst for cross-examination, but rejected Connolly's claim that his conviction must be reversed on that account. Using a state law standard analogous to the federal standard for harmless error and reading any potential waiver problem in Connolly's favor, the SJC reasoned that any error "was harmless beyond a reasonable doubt" because the jury had ample evidence aside from the lab certificates from which it could have concluded that the quantity of cocaine was between 100 and 200 grams. Id. at 374-76; cf. Chapman v. California, 386 U.S. 18, 24, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967) (explaining that federal constitutional error is harmless only if the reviewing court can "declare a belief that it was harmless beyond a reasonable doubt"). It likewise rejected Connolly's other claims of error and affirmed the conviction.

Connolly petitioned for habeas review in federal district court under 28 U.S.C. § 2254. He argued that the state proceedings had violated, inter alia, his Sixth Amendment rights under Melendez-Diaz.4 For the first time, he argued that the prejudice came from the actual admission of the certificate without the analyst's live testimony (rather than from the contemporaneous jury instruction).5 In a thorough opinion, the federal district court denied Connolly's petition. The court, applying the Brecht standard, concluded that "the erroneous admission of the chemical analyses did not have a substantial and injurious effect on the verdict with respect to the weight of cocaine charged." Connolly, 2013 WL 139702, at *3. The court reasoned that although the jury likely did rely on the certificates, the significant amount of other evidence, "including especially the jurors' opportunity to directly examine the cocaine itself," meant that any injury from admitting the certificates was "not substantial enough to allow habeas relief under Brecht." Id.

II.

We review the district court's denial of a petition for a writ of habeas corpus de novo where, as here, the court has taken no evidence and has not made its own factual findings. See, e.g., Kirwan v. Spencer, 631 F.3d 582, 587 (1st Cir.2011).

A.

We begin by describing the applicable legal standards regarding habeas and constitutional error. In 1967, the Supreme Court in Chapman v. California articulated the constitutional harmless error standard, which provides that, on direct appellate review, an error at trial affecting the defendant's constitutional rights will be deemed harmless only if it can be shown to be harmless beyond a reasonable doubt. 386 U.S. at 24, 87 S.Ct. 824. The Court has also repeatedly made clear, however, that the standards that apply on collateral review may differ from those that would apply on direct review. See Brecht, 507 U.S. at 634, 113 S.Ct. 1710. To that end, the Court has held that a federal court on collateral review of a state appellate court's application of Chapman should not apply the same harmless error standard but instead should use an "actual prejudice" standard. Id. at 637, 113 S.Ct. 1710 (quoting United States v. Lane, 474 U.S. 438, 449, 106 S.Ct. 725, 88 L.Ed.2d 814 (1986)) (internal quotation marks omitted). Specifically, as the Court explained in Brecht v. Abrahamson, a habeas petitioner in such a case must show that the error "had substantial and injurious effect or influence in determining the jury's verdict." Id. (quoting Kotteakos v. United States, 328 U.S. 750, 776, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946)) (internal quotation marks omitted).6 That was the law before Congress intervened.

In 1996, Congress passed AEDPA, which provides that a habeas petition may not be granted unless the state court's decision "was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States." 28 U.S.C. § 2254(d)(1). In Mitchell v. Esparza, 540 U.S. 12, 18-19, 124 S.Ct. 7, 157 L.Ed.2d 263 (2003) (per curiam), the Court applied the AEDPA standard to a state court's use of Chapman, explaining that "when a state court determines that a constitutional violation is harmless [under Chapman], a federal court may not award habeas relief under § 2254 unless the harmlessness determination itself was unreasonable." Fry, 551 U.S. at 119, 127 S.Ct. 2321 (describing Mitchell).

Four years later, the Court returned to this field in Fry v. Pliler. In Fry, the state courts had not applied Chapman to determine whether a constitutional error at trial was harmless. See id. at 115, 127 S.Ct. 2321. Without overruling Esparza, the Court ruled that the Brecht standard still governs, regardless of whether the state court applied Chapman. Id. at 119, 127 S.Ct. 2321. As the Court explained, the combined AEDPA/Chapman standard announced in Esparza, "which requires only that the state court's harmless-beyond-a-reasonable-doubt determination be unreasonable," is "more liberal" than the Brecht "actual prejudice" standard. Id. at 119-20, 127 S.Ct. 2321. As a result, the Court held, the Brecht standard "obviously subsumes" the Esparza standard, and federal courts need not formally apply both tests; the Brecht test alone is sufficient. Id. at 120, 127 S.Ct. 2321.

The Courts of Appeals have differed in their interpretations of Fry. Some have concluded that Fry bars the use of the Esparza standard entirely. See Wood v. Ercole, 644 F.3d 83, 94 (2d Cir.2011) ("[W]e conclude that the `unreasonable application of Chapman' standard does not survive Fry."). Others have concluded that the Supreme Court mandated a two-part test, under which a habeas petitioner must first show that the state court's application of Chapman was unreasonable under AEDPA and then must show that the state proceedings had "substantial and injurious effect or influence" under Brecht. See Johnson v. Acevedo, 572 F.3d 398, 404 (7th Cir.2009) ("If the state court has conducted a harmless-error analysis, the federal court must decide whether that analysis was a reasonable application of the Chapman standard. If the answer is yes, then the federal case is over and no collateral relief issues.... If the answer is no — either because the state court never conducted a harmless-error analysis, or because it applied Chapman unreasonably — then § 2254(d) drops out of the picture and the federal court must ... apply the Brecht standard to determine whether the error was harmless." (emphasis added)); see also Gongora v. Thaler, 710 F.3d 267, 274 (5th Cir.2013) (per curiam) (contemplating two-step process in dicta), reh'g en banc denied, 726 F.3d 701, cert. denied, ___ U.S. ___, 134 S.Ct. 941, 187 L.Ed.2d 834 (2014).

Still other circuits have adopted a flexible approach, explaining that "a habeas court remains free" to apply the Esparza test and end the case if the state court's decision was reasonable, but emphasizing that the court need not do that and may instead "go straight to Brecht with full confidence that the AEDPA's stringent standards will also be satisfied." Ruelas v. Wolfenbarger, 580 F.3d 403, 413 (6th Cir.2009).7 This circuit has not yet weighed in on the issue.8

We agree with the Sixth Circuit's flexible approach. As that court noted, Esparza has not been overruled; consequently, we do not read Fry to bar the use of the Esparza test entirely. Contra Wood, 644 F.3d at 94. At the same time, Fry does explicitly state that "it certainly makes no sense to require formal application of both tests." Fry, 551 U.S. at 120, 127 S.Ct. 2321 (first emphasis added). Thus, we do not follow those circuits that require a two-step process. Instead, we note that Fry established that the Brecht standard is harder to satisfy than the Esparza standard, and that any case where the petitioner can show prejudice under the Brecht standard will necessarily meet the Esparza /AEDPA standard. There is clear logic to that position: if an error had a "substantial and injurious" effect on a jury's verdict (Brecht standard), then it is necessarily unreasonable to conclude that the error was harmless beyond a reasonable doubt (Esparza standard).9 See Ruelas, 580 F.3d at 412-13.

Given that logical framework, we conclude that when a state court decides that a constitutional error is harmless beyond a reasonable doubt under Chapman, a federal court on habeas review may choose between two equally valid options. The court may apply Esparza and then move on to the Brecht test only if the state court's decision was an unreasonable application of Chapman. Alternatively, the court may begin with the Brecht test directly.10

B.

With this framework in mind, we turn to the facts of this case. This case lends itself to the first approach, because the petitioner has not met his burden to make the threshold showing that the SJC's conclusion that any error was harmless under Chapman11 was unreasonable.

The error asserted, to be clear, is not that the lab reports contained factual inaccuracies or were unfairly prejudicial to the petitioner. Rather, Connolly asserts that the constitutional error was in denying him the opportunity to cross-examine the lab technician who prepared the drug certificates. In fact, Connolly's brief to the SJC did not argue that the Melendez-Diaz error prejudiced him as to the strength of the evidence of the weight of cocaine. Instead, he focused on the instruction given with the admission of the certificates explaining that, under a state statute, an analyst was not needed to appear in person. This, he argued, prejudiced his ability to argue that the absence of the analyst was a weakness in the prosecution's case. He did not argue to the SJC, nor does he now argue to us, that had he been able to confront the chemist, the outcome would have been different.

Nonetheless, the SJC did consider whether the admission of the drug certificates themselves, not just the improper jury instruction, prejudiced Connolly's case. The SJC explained its reasoning as follows:

Connolly, 913 N.E.2d at 375-76.

Given this claim of error and this analysis as to both drug type and weight, we cannot say that the SJC's conclusion was an unreasonable application of Chapman. "[A]n `unreasonable application of' [Chapman] must be `objectively unreasonable,' not merely wrong; even `clear error' will not suffice.'" White v. Woodall, ___ U.S. ___, 134 S.Ct. 1697, 1702, 188 L.Ed.2d 698 (2014) (quoting Lockyer v. Andrade, 538 U.S. 63, 75-76, 123 S.Ct. 1166, 155 L.Ed.2d 144 (2003)). The harmless error inquiry depends significantly on the facts and is guided largely by the strength of the parties' cases. See, e.g., United States v. Cabrera-Rivera, 583 F.3d 26, 36 (1st Cir.2009); United States v. Carpenter, 403 F.3d 9, 12 (1st Cir.2005).

Here, the government's case was supported by strong evidence apart from the lab certificates. The testimony of highly experienced police officers described the drugs in plain terms that a jury could easily understand — a "large ball, hard," that was "bigger than a baseball" and roughly similar in appearance to a "baked potato." Even without the officers' descriptions, the jurors were told the conversion between grams and ounces; jurors know how much four ounces is and were free to accept that 100 grams, the statutory amount, was less than four ounces. They then could judge for themselves. The jury also saw the hard ball of crack cocaine, though it was more flaked and less solid than it had been when originally found, in the jury room. To be sure, the lab certificates were the most direct way of determining the weight of the drugs, but the jury had obvious alternative methods available for determining the drug weight.12 Melendez-Diaz itself left open the possibility that, when other means of proof aside from the improper certificate exist, the admission of the certificate may be harmless error. 557 U.S. at 329 n. 14, 129 S.Ct. 2527.

We also note that Connolly's trial counsel never requested that the chemist be produced to be cross-examined as to the weight of the substance, suggesting he did not think such a line of inquiry would be productive (or even that he thought it might harm his client).13 That point is reinforced by his failure to make such an argument to the SJC. Indeed, his trial strategy was to challenge identity and the reliability of the police methods used, not to contest the drug type or quantity. These facts are relevant to our assessment of whether there is Brecht or Esparza error. Cf. Neder v. United States, 527 U.S. 1, 16-17, 119 S.Ct. 1827, 144 L.Ed.2d 35 (1999) (holding that failure to instruct jury on one element of crime was harmless beyond a reasonable doubt when overwhelming evidence for that element existed and defendant had not challenged that evidence or argued against that element at trial).

The Massachusetts courts have not been shy about finding reversible error in cases involving Melendez-Diaz error where the facts were different and jurors did not have such other means to determine drug weight or identity. See, e.g., Commonwealth v. Montoya, 464 Mass. 566, 984 N.E.2d 793, 800-01 (2013) (finding non-harmless error where no independent testimony about drug weight in relatable terms was given and evidence of jury confusion about drug weights existed); Commonwealth v. Rivera, 76 Mass.App.Ct. 67, 918 N.E.2d 871, 874 (2009) (finding non-harmless error where, in "borderline" case, jury would be required to apply metric system without testimony about metric unit conversions); Commonwealth v. Rodriguez, 75 Mass.App.Ct. 235, 913 N.E.2d 880, 887 (2009) (finding non-harmless error where jury's only method of determining weight without relying on certificates would be "guesswork" on a "close question"). Particularly in light of these distinguishable cases, the facts of this case show that the SJC's determination that the error was harmless was not unreasonable, regardless of whether we might have found the error harmless on direct review. Cf. White, 134 S.Ct. at 1702 ("[A] state prisoner must show that the state court's ruling on the claim being presented in federal court was so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement." (quoting Harrington v. Richter, ___ U.S. ___, 131 S.Ct. 770, 786-87, 178 L.Ed.2d 624 (2011)) (internal quotation mark omitted)).

Under Fry, we may stop here, as petitioner is not entitled to habeas relief under the Esparza standard. Nonetheless, we briefly explain why, even if the SJC's application of a Chapman harmless error test was unreasonable, petitioner still would not be entitled to habeas relief because he cannot show error under the Brecht standard. We return to the nature of the claimed prejudice from the error.

As noted, Connolly has not challenged the accuracy of the lab certificates. Rather, his claim is limited to the deprivation of his opportunity to cross-examine the lab analyst, and the erroneous jury instruction on that topic. Without a meaningful challenge to the actual contents of the certificates, however, Connolly cannot show substantial and injurious effects on the jury's verdict. See Dominguez v. Duval, 527 Fed.Appx. 38, 41 (1st Cir.2013) ("The argument for prejudice ... addresses not the contents of the report but the absence of the examining pathologist.... But not only is the substance of the pathologist's hypothetical testimony a matter of pure speculation, the possibility that any such testimony would have swayed the jury toward accepting [petitioner's] account is downright unrealistic."). Connolly does not claim that, had he been able to cross-examine the lab technician, the jury would have been more likely to conclude that the ball of cocaine weighed under 100 grams.14 Without any such claim, he cannot show on habeas review that the admission of the drug certificates had a "substantial and injurious" effect on the jury's decision. And even if he made that claim, he has not put forward any evidence indicating what would have been revealed on cross-examination, leaving the support for his claim a matter of "pure speculation." Id. That is insufficient to show a "substantial and injurious" effect on the verdict.

III.

We affirm.

THOMPSON, Circuit Judge, concurring.

This case aptly illustrates the limited nature of our habeas review even in the face of an admitted constitutional error. The myriad obstacles to habeas review, especially the barricades thrown up and reinforced by Supreme Court precedent, compel us to affirm Connolly's conviction. Although I fully concur in the Court's judgment, I write separately to flesh out why we are powerless to intervene.

Put simply, the Supreme Court has recently reiterated that although some federal judges find the scope of habeas review to be too limited, our authority to grant the writ is circumscribed by 28 U.S.C. § 2254, a statute we are bound to obey. White v. Woodall, ___ U.S. ___, 134 S.Ct. 1697, 1701-02, 188 L.Ed.2d 698 (2014). We may only grant habeas relief where there has been a misapplication of "clearly established Federal law" of sufficient severity. Thaler v. Haynes, 559 U.S. 43, 47, 130 S.Ct. 1171, 175 L.Ed.2d 1003 (2010) (per curiam) (internal quotation marks omitted). Significantly, the relevant federal law must have been clearly established as of the time of the state court decision, meaning that subsequent developments are not relevant to habeas review. Williams v. Taylor, 529 U.S. 362, 412, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000). Furthermore, for us to find a state court unreasonably applied clearly established federal law, its application "must be `objectively unreasonable,' not merely wrong; even `clear error' will not suffice." White, 134 S.Ct. at 1702, (quoting Lockyer v. Andrade, 538 U.S. 63, 75-76, 123 S.Ct. 1166, 155 L.Ed.2d 144 (2003)).15

Everyone agrees a Melendez-Diaz error occurred at Connolly's trial, which — as the majority opinion discusses — requires us to engage in a harmless error analysis. The SJC determined in Connolly's direct appeal that the jury could find beyond a reasonable doubt that the drugs weighed over 100 grams because the jury had the actual drugs with them as they deliberated. Thus, the SJC concluded the Melendez-Diaz error was harmless. See Commonwealth v. Connolly, 454 Mass. 808, 913 N.E.2d 356, 375-76 (2009).

This limits the scope of our habeas review to determining whether or not the SJC's finding of harmless error constituted an unreasonable application of clearly established federal law. In other words, Connolly needs to demonstrate that it has been clearly established as a matter of federal law that a lay jury may not make a finding of drug weight based on the opportunity to observe and handle those drugs. This is where our review rushes headlong into the habeas wall.

Timing is everything, and unfortunately time was not on Connolly's side. The Massachusetts appellate courts appear to have backed away from the SJC's holding in Connolly. Subsequent cases have distinguished it and concluded a Melendez-Diaz error with respect to drug certificates was not harmless beyond a reasonable doubt. Just over a week after the SJC handed down Connolly, the Appeals Court found that the erroneous admission of drug certificates was not harmless where the alleged amount of drugs exceeded the statutory threshold of 100 grams by 42.5 games, and were spread across eight bags, one of which contained 106.5 grams of drugs. See Commonwealth v. Rodriguez, 75 Mass.App.Ct. 235, 913 N.E.2d 880, 887 (2009). The Appeals Court distinguished Connolly and reasoned that in the absence of the drug certificates, "determining the weights of the amounts at issue, particularly whether they were over or under 100 grams, would involve too much guesswork on too close a question in these circumstances." Id. (citing Connolly, 913 N.E.2d at 376).16 The Appeals Court has gone on to find non-harmless error in several opinions issued subsequent to Connolly and Rodriguez. See Commonwealth v. Rivera, 76 Mass.App.Ct. 67, 918 N.E.2d 871, 874-75 (2009) (admission of certificates not harmless error where multiple bags of heroin and cocaine were all within 100 grams of the statutory thresholds); Commonwealth v. DePina, 75 Mass.App.Ct. 842, 917 N.E.2d 781, 789-90 (2009) (finding nonharmless error where alleged amount of drugs exceeded statutory threshold of 14 grams by 0.24 grams and was contained in 8 separate bags).

The SJC itself distinguished Connolly in Commonwealth v. Montoya, 464 Mass. 566, 984 N.E.2d 793 (2013). Montoya contrasted the evidence in Connolly, "a ball of solid cocaine that looked like a baked potato [and] was bigger than a baseball," with the drugs in that case, 39.74 grams (which exceeded the threshold by approximately 11.8 grams) "packaged in twenty individual bags." Id. at 800-01 (internal quotation marks omitted). The SJC went on to find "[t]hat jurors had the opportunity to handle this evidence in the jury room, without instruments or any objects of known weight for comparison, does not render the admission of the drug certificates harmless beyond a reasonable doubt," and cited the Appeal's Courts decisions in Rodriguez, Rivera, and DePina with approval. Id. at 801.

Thus, in Montoya it appears that the SJC considered the fact that the drugs involved in that case were spread across multiple packages to be the key distinguishing feature between its decision there and its holding in Connolly. It is at least debatable whether or not the SJC would come to the same result if presented with Connolly's appeal today in light, especially in light of its approval of Rodriguez, where one of the bags of drugs weighed more than 100 grams.

Nevertheless, the subsequent development and potential uncertainty with respect to a matter of state law are simply irrelevant to our habeas review in this instance. Connolly has not directed this Court's attention to any Supreme Court precedent clearly establishing that a lay jury may not — consistent with the federal constitution — make the drug weight determination on its own. My own research has failed to locate any Supreme Court precedents in this area either.17 As the Supreme Court has reminded us, the absence of Supreme Court holdings inevitably means that there is no "clearly established federal law." Thus, Connolly's habeas petition is doomed. Under the strict habeas standard, it matters not whether the SJC might come to a different conclusion today or whether we would have arrived at a different result had this case reached us on direct appeal. In light of these restrictions, we have no choice but to dismiss Connolly's habeas petition.