GAYNOR, J.:

The usual or average period of human gestation, i.e., from the beginning of pregnancy to the maturity of the fœtus, when delivery should and usually does take place, is 10 lunar months, or 40 weeks, i.e., 280 days. The period from intercourse to delivery may not be identical with that of gestation, but much longer. There is much literature, ancient and modern, in favor of the claim that this latter period may normally last 11 calendar months, especially to save a widow and her tardy offspring from anything worse than suspicion; but it has scarcely survived the immortal satire of Rabelais (B'k 1, ch. 3). Pregnancy does not always take place and gestation therefore begin at the time of intercourse but several days (even more than a week, respectable medical writers say) later, when the ovum has become ripe to receive the male seed, which retains its vitality in the generative tract meanwhile. By making allowance for this in reckoning back from the birth of a child, and also for the several days, at least, that delivery may be postponed beyond the usual period of gestation, i.e., after the growth of the fœtus has become complete, from 10 to 20 days may be easily added (Coke on Lit. by H. B., p. 123 b., n. 1 2; 2 Greenl. on Ev. sec. 152; Bouv. L.D. vol. 1 [Rawl. Rev.], p. 882; Herold on Legal Med. chap. 40; Lusk on Midwifery, p. 108; Am. Text Book of Obstetrics, p. 176). Hence 300 days from the time of last possible access is generally established by statute on the continent of Europe as the period to determine legitimacy (Code Napoleon, art. 312). In this State a whole year has long been established by statute as the period in proceedings for the support of bastards (1 R.S. p. 641; Code Crim. Pro. sec. 838).

The appellant was born only 241 days after his father's death. But it is claimed that his father did not have access to his mother for 3 calendar months and 17 days preceding his death. This claim, if true, would lengthen the period from the latest possible time of intercourse to the time of delivery 108 days more, i.e., to 349 days, and present a serious case of legitimacy. It rests on the testimony of the mother of the two co tenants of the appellant, who are contesting his legitimacy. She testified that his father lived in the house of his brother, then the husband of her sister, and afterwards her husband and the father of her said children, for "about two months" before he was taken to the hospital, where sexual access to his wife was impossible, and where he died of phthisis after about six weeks (47 days); that she also lived in her said sister's family at the same time, and helped to take care of him; that he never went out while living there, and that his wife never came there to visit him. This is the testimony of a biased and hostile witness, and also improbable in itself. The testimony is that he was brought by his brother to the latter's house from the house where he and his wife were living together (evidently in poverty owing to his sickness), for better care. No estrangement existed between them, nor any reason why his wife should abandon him, so far as appears. And the length of time is left in uncertainty. The witness, testifying from unaided memory after the lapse of over twenty-five years, gives it as "about" two months. We all know how natural it is for a biased witness to stretch weeks into months, even without intending to falsify, especially after such a lapse of time. Except for this testimony, and some slight corroboration by a sister of the witness who visited the house, and testified that the time was "about" two months, the legitimacy of the appellant is not even blown upon.

The presumption of fact of legitimacy is one of the strongest known to the law, and of course it cannot be overthrown except by evidence which is stronger. The burden of proof is on the party asserting illegitimacy, and the rule in a case like the present one has been declared to be "that to bastardize the issue of a married woman, it must be shown beyond all reasonable doubt that there was no such access as could have enabled the husband to be the father of the child" ( Cross v. Cross, 3 Paige, 139; Van Aernam v. Van Aernam, 1 Barb. Ch. 375; Caujolle v. Ferrié, 23 N.Y. 90; Matter of Matthews, 153 N.Y. 443; Br. Had. Com. Am. Ed. p. 384, n, and cases cited; Lawson on Presumpt. Ev. p. 108). The feeble and indefinite evidence relied on here is evidently insufficient. If 280 days be taken as the period of gestation, there were within it 39 days of possible sexual access of the appellant's parents before his father went to the hospital, and if 300 days be taken as the period allowed from the latest opportunity of access to delivery, there were 59. The father could therefore have been at his brother's house without access to his wife for several weeks — for nearly six weeks — before going to the hospital without the legitimacy of the appellant being jeopardized. Who will say that even an honest recollection of two months after the lapse of nearly twenty-six years may not be just as safely taken for five weeks, or even much less?

The hospital doctor testified that he could say that during the last two months of life of a man in the father's state of health it was improbable but not impossible that he could beget a child. Not only was this evidence far from conclusive, but it left out the previous month and 17 days.

That the appellant's mother did not make known his birth to her deceased husband's kin is of no significance, for they had ignored her from the beginning. They never met her except at her husband's grave on the day of the funeral, and they left her there. The appellant did go to his grandfather and his uncle in his boyhood and announce his kinship.

The judgment should be reversed.

HIRSCHBERG, P.J., HOOKER, RICH and MILLER, JJ., concurred.

Judgment reversed and new trial granted, costs to abide the final award of costs.