Judgment entered June 9, 1967 after a jury trial in favor of plaintiffs, reversed on the facts and on the law and in the exercise of discretion, and a new trial directed, with costs and disbursements to abide the event. We conclude that the defendants were deprived of a fair trial and an unprejudiced consideration of the ease by reason of the continuing exchanges between court and counsel for the defendants in the presence of the jury. Although defendants’ counsel was not free from fault in this area, some of the observations of the court may well have prejudiced the defendants. There were criticisms of the defendants’ counsel to such an extent that in our opinion the calm, dispassionate and deliberate consideration of the facts by the jury was unduly impeded. The Trial Judge should at all times maintain an impartial forbearance. The development of the facts in the presence of the jury should be uncomplicated by personalities *913and acrimony. (Habenicht v. B.N.O. Theatres¡ 23 A D 2d 378; Salzano V. City of New York, 22 A D 2d 656; Levy v. Reilly, 18 A D 2d 632; Buckley v. 2570 Broadway Gorp., 12 A D 2d 473; Kamen Soap Prods. Go. v. Prusansky & Prusansky, 11 A D 2d 676.) We must note our disagreement with the concurring opinion insofar as error is charged in the refusal to instruct the jury as to the unexplained failure of the plaintiff’s physician to testify. “ The courts in considering the failure of a party to call a witness within his control by being in his employ or so related to him or on such terms with him as to indicate friendliness, have stated that it gives rise to a presumption or warrants an inference by the jury that the testimony would have been adverse to the party so failing to call the witness or would have been unfavorable or prejudicial to him; but this has been usually qualified by stating what is plainly the correct rule for instructing a jury, viz., that in such circumstances the jury have a right to acoept the testimony before them which might have been but was not controverted, and to take it most strongly against the party who might have controverted it but failed to do so.” (Perlman v. Shanck, 192 App. Div. 179, 183, italics in original.) The doctor who administered the drug did not fall within the defined class of witnesses. Hence the court’s refusal to charge as to his nonappearance as a witness did not constitute error. (Vollmer v. Automobile Fire Ins. Co., 207 App. Div. 67, 70; Beehil v. Fraas, 129 App. Div. 563, 566, revd. on other grounds 197 1ST. Y. 64; 2 Wigmore, Evidence [3d ed.], § 287, p. 168; Richardson, Evidence [8th ed.], § 92, p. 65; McCormick, Evidence, § 249, p. 534.) Concur — Capozzoli, J. P., Markewich and Tilzer, JJ.; McGivern, J. concurs in part and dissents in part in the following memorandum: Although I concur in the direction of a new trial, it is not alone for the sole reason assigned by the majority. The conduct of a trial should indeed be seriously viewed, but the benefit of a new trial should not for improper conduct alone be given when as in the instant case, the medical support of the plaintiffs’ ease is so weak, or nonexistent, the complaint could as justifiably be dismissed on the record before us now. Two of the plaintiffs’ doctors never categorically said that the drug Altafur was the competent producing cause of her malady. Dr. Jacobson did, but on cross-examination, even he admitted he was not familiar with the proper dosage, had never used the drug and had never seen the instructions on the package. Yet, he was the plaintiffs’ “ expert ” witness. And not one of the physicians called by her actually treated her at the time in suit. Taking an informed notice of the possible side effects of all antibiotics, not excluding aspirin, the effort to hold the defendant liable for an overdosage of one drug, when not less than twelve others were also given, stretches the doctrine of proximate cause to the breaking point. Particularly is this so when it is admitted that plaintiff received not only an overdosage of Altafur but before that received massive doses of “ Stilbesterol, Dermerol, Gantrisin, aspirin, furadantin, seconal, aureomycin, Chloromycetin, ilosone, paregoric, mereuhydrin, erythromycin.” In any event, a new trial, eleven years after the administration of the drug Altafur, in itself borders on a travesty. The record discloses not only the inherent weakness of the plaintiffs’ case but also many errors of an extraordinary character warranting a new trial. These errors, briefed by both parties on this appeal, necessitate a disposition and should not be ignored by an appellate court ordering a new trial. There was the arbitrary abridgment and preclusion of critical cross-examination, and the exclusion of evidence that the cause of the plaintiff wife’s visual disability was her own physical condition and the gigantic administration of antibiotics other than Altafur properly given. Then, in addition, there were two crucial errors: (a) the refusal to grant the defendant’s request to charge *914that the failure of the plaintiffs to call the plaintiff wife’s own doctor, who recommended the overdose of Altafur, was a factor properly to be considered by the jury, and (b) the refusal of the court to charge the jury in respect of the effect of an overdose of Altafur in contravention of the instructions for safe usage contained in the package insert. Concerning the failure of the plaintiffs to call Dr. Shapiro, I completely dissent from the majority’s conclusion that “ the court’s refusal to charge as to his nonappearance as a witness did not constitute error.” In my view, Dr. Shapiro, as the plaintiff wife’s treating physician, the one responsible for and who recommended the overdosage, was the one witness most likely to shed light on her condition. He should have been produced, or his absence explained, out of the very necessity of the case. It was with him the plaintiff wife had a special relationship^ and it was both natural and reasonable to expect his testimony, a trial being a quest for the truth. In the nature of things he had a superior and a unique knowledge of a material point on which the plaintiffs had the 'burden of proof. Additionally, the record before us is open to the interpretation that the administration of the drug in question may have helped save the plaintiff wife’s life, but the overdose, if any, as prescribed by her own doctor, may have triggered the consequences to her eyes, which constituted the main gravamen of her complaint. Her own expert, Dr. Jacobson, admitted her life was in danger, and there is testimony that the administration of Altafur saved her life. Norwich, the defendant-appellant, meritoriously argues that her physician may have intentionally administered overdosages to rescue her from the brink. “ The art of healing frequently calls for a balancing of risks and dangers to a patient”. (Perlmutter v. Beth David Hosp., 308 N. Y. 100, 107.) If so, this might absolve her physician and affect the issue of proximate cause as to the manufacturer. Thus, the testimony of the treating physician, who prescribed the dosages, became a vital factor in her cause and in an understanding of the entire case. A charge, appropriately phrased, on the subject of his unexplained failure to testify, was in order and warranted under the proof and circumstances of the case. A request of this nature was made. It was refused by the court. In my judgment, this was error. Actually, the charge requested was the following: “VII. If you find that the plaintiff’s obstetrician, Dr. Shapiro, who administered the drug Altafur to the plaintiff was in a position to give evidence supporting the claims made by the plaintiff in this law suit or to contradict evidence offered by the defendant in this case and that the plaintiff has not shown reasonable explanation for the failure to call or subpoena this doctor to court you may infer that the testimony of said Dr. Shapiro would not support the plaintiffs’ version of the case and draw the strongest inference against the plaintiffs that the opposing evidence permits. Refused. D. W. T., J. S. C.” In my judgment, the foregoing request was unexceptionable and is countenanced, in the very specifics of its language, by Laffin v. Ryan (4 A D 2d 21, 24) and Noce v. Kaufman (2 N Y 2d 347, 353). There are other cases of like tenor: (Kane v. Rochester Ry. Co., 74 App. Div. 575, 580; Minch v. New York & Queens County Ry. Co., 80 App. Div. 324; Deutschman v. Third Ave. R. R. Co., 87 App. Div. 503, 513; Brotherton v. Barber Asphalt Paving Co., 117 App. Div. 791 Reehil v. Fraas, 129 App. Div. 563, 566; Perlman v. Shanck, 192 App. Div. 179; Sabowska v. Coney Is. & Brooklyn R. R. Co., 174 App. Div. 913; Group v. Szenher, 260 App. Div. 308; Masterson v. Solomon, 191 Misc. 635; Seligson, Morris & Neuburger v. Fairbanks Whitney Corp., 22 A D 2d 625; Goldman v. State of New York, 28 A D 2d 782; Meditz v. Liggett & Myers Tobacco Co., 167 Misc. 176, 181, affd. 25 N. Y. S. 2d 315; Falcone v. Repetti, 61 Misc 2d 407), In the Seligson, Morris case (supra, p. 630) this *915court, per Breitel, J., said: “ This rule is applied, for example, where a plaintiff fails to call his own physician as a witness”. Other jurisdictions apply the rule with vigor. To cite only a few, and involving physicians: (Secondino v. New Haven Gas Co., 147 Conn. 672; Ferne v. Chadderton, 363 Pa. 191, 198; Blackstone v. Osche, 192 F. Supp. 174; Donet v. Prudential Ins. Co. of America, 23 S. W. 2d 1104 [Mo.]; Anders v. Employers Liab. Assur. Corp., 50 So. 2d 87 [La. App.]; Phillips v. Reece, 106 Ga. App. 779; Feldstein v. Harrington, 4 Wis. 2d 380; Shockman v. Union Transfer Co., 220 Minn. 334; Craig v. Marquette Cement Mfg. Co., 190 Tenn. 234; Great American Ind. Co. v. Mayer, 272 S W 2d 569 [Tex. Civ. App.]; American Employers Ins. Co. v. Kellum, 185 S W 2d 113 [Tex. Civ. App.]). In the latter ease, the court held the jury had a right to consider the doctor’s absence even when the doctor told the plaintiff ” if he was made to come to court he would not do the plaintiff any good ”. To the same effect are the leading text writers. (See 2 Wigmore, Evidence [3d ed.], § 285, p. 162; McCormick, Evidence, § 249, p. 534; 2 Ford, Evidence, § 124; Fisch, New York Evidence, §§ 556, 1126; see, also, Schwartz, Trial of Automobile Accident Cases, [2d ed.] § 390, p. 430; Rule of Evidence in Negligence Cases, Kramer [2d ed.], Practicing Law Institute [1960], p. 72; 5 ALR 2d 920; 31A C. J. S., Evidence, § 156[3].) Tested by these principles, the defendant’s request should have been granted. Thus, I am completely nonplussed by the majority’s pronunciemento: “ The doctor who administered the drug did not fall within the defined class of witnesses”. By this the majority become the exponents of a most novel and utterly untenable proposition, hitherto unknown to the law. And, in my view, the citations on which they rely, are all to the contrary. Perlman v. Shanck (192 App. Div. 179, 184-185, supra) involved a doctor who resided in Toronto, Ontario, and the court did not reject the rule, but held that generally it should be confined to witnesses within the jurisdiction of the court whose attendance upon the trial might be procured by subpoena, citing Reehil v. Fraas (supra). Furthermore, the court regarded the testimony of the Toronto doctor as only cumulative. Yet, even then, the court pointedly said: “If there were anything improbable in the plaintiff’s testimony which might have been corroborated by the testimony of the physician, perhaps the jury might have been warranted in inferring that the physician would not have corroborated the plaintiff”. Vollmer v. Automobile Fire Ins. Co. (207 App. Div. 67, 70) cited by the majority, has no value except a quotation from the Perlman case (supra). Reehil v. Fraas (129 App. Div. 563, 566 supra) also cited by the majority, is a landmark case supporting the defendant’s request. Noteworthy is the statement “ It does not apply to all such witnesses, but only to those of whom it may be naturally inferred that they are of good will to the party, such as his business partners or associates, his employees, his physician, and the like”. (Italics supplied.) As for the majority’s citation of McCormick on Evidence, I embrace the quotation at page 534: “ It is generally agreed that when a potential witness is available, and appears to have special information relevant to the ease, so that his testimony would not merely be cumulative, and where his relationship with one of the parties is such that the witness would ordinarily be expected to favor him, then if such party does not produce his testimony, the inference arises that it would have been unfavorable.” And footnote 8 reads: “ For example, if the potential witness is the family doctor of the party.” (Italics added.) Further, although there was expert testimony that an overdose of the drug could be harmful, as also testimony that the dosage prescribed was continued despite an early complaint of eye “flashes”, a charge on this subject of overdose was requested and denied. *916This too was reversible error. (4 Weinstein-Korn-Miller, N. Y. Civ. Prac., par. 4404.18. This subject was of crucial importance on the issue of proximate cause, but it was only obliquely and ambiguously mentioned by the court in that portion of the charge relating to ivarranty, the court using the words “ without any misuse or mishandling of the drug ”, the court failing to specify whether by the patient or by her doctors. The essentiality of proximate cause in a warranty case has been stressed. See Natale v. Pepsi- Cola Co. (7 A D 2d 282, 284) wherein it is noted: “ If the essential cause of the occurrence was something other than the breach of the implied warranty, no matter what, the defendant is not liable. * * * It is the breach of the implied warranty of merchantibility which gives rise to the liability and the evidence must affirmatively establish that it is causally related to the alleged injuries ”. Explicitly, the court in the case at bar, failed entirely to charge on the specific subject of “ overdose ” in contravention of the indicated permissible dosage and as a possible cause of the plaintiff wife’s malady. In a close situation such as this, involving a narrow but weighty issue, this inadequacy in the charge becomes fatal. Basic inadequacy or confusion in a charge preclude a fair consideration by a jury and require the reversal of a judgment and a new trial in the interests of justice. (Bacon v. Celeste, 30 A D 2d 324; Molnar v. Slattery Contr. Co., 8 A D 2d 95, 100.) Accordingly, although I take a dim view of the plaintiffs’ case, I conclude the foregoing errors mandate a new trial, with appropriate direction and guidance to the trial court in accordance with the views expressed herein.