Rickerson was charged with criminal sexual penetration of a minor. At trial the jury was instructed on both the crimes of criminal sexual penetration of a minor and a lesser included offense of criminal sexual *667contact of a minor. After several hours of deliberation and four or five votes the jury informed the court that they could not agree on a verdict. The judge reassembled the jurors in the courtroom and asked the foreman to tell him, without divulging whether the vote was for or against conviction, what the numerical results of the voting had been. He was informed that the first vote on the greater charge of criminal sexual penetration had been three to nine and that it had changed only one vote over the course of the deliberations. The judge was also informed that the jury did not feel they could reach a unanimous verdict on the lesser included offense of criminal sexual contact. The judge, speaking to the jury, said “the court is going to let you retire to the jury room again; and that’s about all I want to say at this time. The jury will retire.” Rickerson moved for a mistrial based on the court’s inquiry into the numerical split of the jury and its statement to the jury, arguing that it was in effect a modified shotgun instruction. The motion was denied. The jury returned after two hours of additional deliberation with a guilty verdict on the lesser included charge of criminal sexual contact, of a minor.
The defendant appealed his conviction to the Court of Appeals, alleging three errors by the trial court. The Court of Appeals reversed. We granted certiorari. The question on certiorari is whether an inquiry by the trial court limited to the numerical division of the jury violates due process. Under the facts of this case we hold that it does not.
This question has been addressed by this Court and the Court of Appeals in several opinions. See State v. McCarter, 93 N.M. 708, 604 P.2d 1242 (1980); State v. Nelson, 63 N.M. 428, 321 P.2d 202 (1958); State v. Turner, 90 N.M. 79, 559 P.2d 1206 (Ct.App.1976), cert. denied, 90 N.M. 9, 558 P.2d 621 (1977); State v. Aragon, 89 N.M. 91, 547 P.2d 574 (Ct.App.1976), cert. denied, 89 N.M. 206, 549 P.2d 284 (1976), and Pirch v. Firestone Tire & Rubber Co., 80 N.M. 323, 455 P.2d 189 (Ct.App. 1969), cert. denied, 80 N.M. 316, 454 P.2d 973 (1969). Prior to Aragon, inquiries into the numerical division of juries were dealt with under the rule established in the cases of Nelson and Pirch. The rule was that convictions would be reversed if the cumulative effect of the trial court’s actions had a coercive effect on the jury.
The factors considered in determining if a court’s inquiry was coercive under the Nelson-Pirch rule were: (a) whether any additional instruction or instructions, especially a shotgun instruction, were given: (b) whether the court failed to caution a jury not to surrender honest convictions, thus pressuring holdout jurors to conform, and (c) whether the court established time limits on further deliberations with the threat of a mistrial. This test was changed by the Court of Appeals in the Aragon case. Although the court in Aragon announced a new test to be used in future cases dealing with inquiry into the numerical division of the jurors, the Aragon case itself was decided based on the Nelson-Pirch test. The rule announced in Aragon was that any inquiry into the numerical division of the jurors is reversible error. There is no need under the Aragon rule to establish whether the inquiry into the numerical division had a coercive effect on the jury. The Court of Appeals based this test on the United States Supreme Court case of Brasfield v. United States, 272 U.S. 448, 47 S.Ct. 135, 71 L.Ed. 345 (1926). The court in Aragon stated: “Because the error [inquiry into the numerical split of the jury] goes to a ‘fair and impartial’ trial, the error violates due process. Accordingly, Brasfield v. United States, supra, applies to New Mexico courts.” 89 N.M. at 97, 547 P.2d at 580.
Only the State of Michigan in addition to New Mexico has held that Bras-field involved constitutional principles and was thus applicable to the states. See People v. Wilson, 390 Mich. 689, 213 N.W.2d 193 (1973). The remainder of the states and *668federal courts that have considered the issue have held that Brasfield was an exercise of the Supreme Court’s supervisory powers over the federal courts and not an announcement of a principle of constitutional law, or they have held that inquiry into the numerical division of a jury is not error per se without discussing Brasfield. As stated by the Eighth Circuit in Cornell v. State of Iowa, 628 F.2d 1044, 1047 (8th Cir. 1980):
When read in the light of the rationale suggested in Burton, we think the rule in Brasfield is more easily understood, not as an announcement of a mandatory principle of substantive constitutional doctrine, but as an administrative admonition to the lower federal courts based upon carefully considered notions of sound judicial practice.
See also Ellis v. Reed, 596 F.2d 1195 (4th Cir. 1979), cert. denied, 444 U.S. 973, 100 S.Ct. 468, 62 L.Ed.2d 388 (1979); People v. Carter, 68 Cal.2d 810, 69 Cal.Rptr. 297, 442 P.2d 353 (1968); Lowe v. People, 175 Colo. 491, 488 P.2d 559 (1971); Muhammad v. State, 243 Ga. 404, 254 S.E.2d 356 (1979); People v. Kirk, 76 Ill.App.3d 459, 31 Ill.Dec. 835, 394 N.E.2d 1212 (1979), cert. denied, 447 U.S. 925, 100 S.Ct. 3019, 65 L.Ed.2d 1118 (1980); State v. Cornell, 266 N.W.2d 15 (Iowa 1978), cert. denied, 439 U.S. 947, 99 S.Ct. 340, 58 L.Ed.2d 338 (1978); State v. Smith, 431 S.W.2d 74 (Mo.1968); see e. g. Gov’t of Virgin Islands v. Romain, 600 F.2d 435 (3rd Cir. 1979). We have reviewed the Brasfield case and hold that it was an exercise of the Supreme Court’s supervisory powers and did not involve substantive constitutional principles. It is, therefore, not binding on the states. See Crist v. Bretz, 437 U.S. 28, 98 S.Ct. 2156, 57 L.Ed.2d 24 (1978); People v. Kirk, supra. Therefore, we reverse that portion of the Aragon case which held to the contrary.
While inquiry into the numerical division of the jury is not to be encouraged, see Cornell v. State of Iowa, supra and Lowe v. People, supra, it is not error per se. We reaffirm the Nelson-Pirch rule that such inquiries are reversible error only when shown to have a coercive effect on the jury. The inquiry itself is not coercive since the jury is already well aware of its numerical split. Ellis v. Reed, supra.
There is some justification for inquiries as to probability of agreement among the jury when done pursuant to the court’s duty to assure that a verdict is reached, People v. Carter, supra, and in determining whether further deliberations are needed or if the jury should be discharged, People v. Hall, 25 Cal.App.2d 336, 77 P.2d 244 (1938); Linscomb v. State, 545 P.2d 1272 (Okla.Crim.App.1976). Such an inquiry may also be necessary to protect the defendant from double jeopardy consequences when more than one count is presented to the jury. See N.M.R.Crim.P. 44(d), N.M.S.A. 1978 (Repl.Pamp. 1980); O’Kelly v. State, 94 N.M. 74, 607 P.2d 612 (1980).
We remand this case to the Court of Appeals for its consideration of other issues raised on the appeal not dealt with in its memorandum opinion.
IT IS SO ORDERED.
EASLEY, C. J., SOSA, Senior Justice, and FEDERICI and RIORDAN, JJ., concur.