The advice of this court is asked concerning the proper construction of the residuary clause of the will of Samuel Hillhouse, late of Wethersfield, deceased, which is as follows: — “All the residue of my estate, real and personal, I give and devise to the following named persons, to be divided equally among them: — my sisters Rachel and Sarah, the grandchildren of my deceased brother William, and the grandchildren of my deceased sisters Delia and Mary; meaning by this to include all said grandchildren living at the time of my decease; provided that the share of Elizabeth Raymond, daughter of my deceased nephew, James H. Raymond, shall be held by John Beach, of Goshen, in trust for her benefit as hereinbefore provided.” And the precise question is, whether the estate should be distributed among the beneficiaries per stirpes, or per capita.
To aid in the construction we invoke, first, the paramount rule that requires u.s to ascertain if possible the intention of the testator from the language of the will and all the circumstances. But unfortunately the language hero employed is considered so doubtful as to be made the basis of directly opposing arguments.
The counsel who contended for a distribution per capita claimed that the testator’s use of the word “persons” was a clear individualization of his beneficiaries. But wo think *473this word, in the connection, may as well refer to a class of persons. The phrase, “ the following named persons,” it is true raises an expectation that the specific names of the beneficiaries are about to be given, but in this case the testator, along with the names of his two living sisters, gives the names of his two deceased sisters and of his deceased brother, whose grandchildren are to receive the legacy. Now did the testator mention the names of the deceased persons only by way of designating the individual grandchildren who were to take, or were the names thus given as the proper heads of the classes who were thus to take ? The latter construction is at least as natural and reasonable-as the former.
The will is silent as to the names and number of the grandchildren; they are an uncertain body to be first ascertained at the time of the testator’s death, and only those then living are included.
Neither do the words, “ to be divided equally among them,” necessarily import a division of the property among individuals, for they apply just as readily and appropriately to a division among classes.
If now we pass from the particular clause under discussion to other parts of the will, we can find nothing inconsistent with a distribution of the residuum per stirpes. If we look at the clause immediately preceding we shall see that when the testator had in mind a per capita distribution of certain property in other states among his nephews and nieces, it was very clearly expressed, and in very different language too from that under consideration. The testator might easily in the clause we are considering have given all the residue, (following the style of the clause just referred to), directly to his two sisters and to all his great nephews and great nieces, but for some reason he preferred to name the five per-'U sons who stood nearest in kinship — two living and three 1 deceased — the latter leaving grandchildren who were to take the estate. In so doing, we think it probable that the testator had in mind a division of the property into five parts. The J will elsewhere discloses the fact that the two living sisters were the especial objects of his bounty, and we think it *474improbable that lie intended to make tlieir shares as small as any one of his nineteen great nephews and great nieces.
But we are confronted with the English rule of construction, adopted in Massachusetts, New York, and some other states, that “a gift to children, grandchildren, or heirs, is equivalent to naming them and is a gift to them individually.” 2 Jarman on Wills, 1st Am. ed., 111. The same author, after citing the above rule, adds, “ But this mode of construction will yield to a very faint glimpse of a different intention in tbe context.”
If the above rule is so easily set aside, it would seem equally reasonable that it should also yield to the presumption in favor of the natural heirs or next of kin, for a distribution according to the statute, in all cases where the language of the will is consistent with such a distribution, and the real intention of the testator is in doubt. And such a rule of ‘construction was adopted by this court in Lyon v. Acker, 33 Conn., 224, and has been recognized in the state of Pennsylvania. In Minter’s Appeal, 40 Penn. S. R., 111, Lowrie, C. J., justly remarked: “When we find a man distributing his estate, in whole or in part, among his next of kin, and he leaves the proportions in which they are to take doubtful, it is quite natural for us to suppose that he had the statutory or customary form of distribution in his mind and to interpret his will accordingly.” And the opinion of the same judge in Fissel’s Appeal, 27 Penn. S. R., 55, is equally pertinent.
In the case at bar it is conceded that under the statute governing the descent of estates, the grandchildren of the testator’s two deceased sisters and deceased brother would take, by right of representation, their deceased parents’' share. ¡The English rule, that referring to children is the same as if they were individually named in the will, in view of the exceptions which so easily set it aside, is of little practical importance, and it does not seem to have been recognized in this state, although cases have been discussed and decided where its application would have been controlling. Our decisions quite uniformly have been in favor of a per stirpes distribution in all cases ahalagous to this.
*475The case of Lyon v. Acker, supra, would seem t decisive of this. The will contained the clause, “I give my three daughters, Mary, Susan, and Josephine, and t. children of my son Samuel, my homestead, to them and then assigns forever, share and share alike.” It was held that the children of Samuel took per stirpes, and not per capita. The cases are parallel, except as to the degree of kindred and the language employed to denote the portion given to each, and the difference in these respects can make no difference in the principles of construction.
There are several other cases, which, though not as controlling as the above, yet support the views here expressed, and we believe there are no decisions in this state inconsistent with these views. In Talcott v. Talcott, 39 Conn., 186, the devise was to the children of one Mrs. Burke, deceased, and though the names of each of the children were given, yet it was held that they took as a class, and not per capita. In Bond’s Appeal from Probate, 31 Conn., 183, the devise was, “I give to my children and their heirs respectively, to be divided in equal shares between them.” At the date of the will and at his death the testator had four children living, and four others had previously died, all leaving children. No reason appeared for supposing that the testator had any preference for his surviving children over these grandchildren, and it was held that the estate in question was to be distributed in equal shares among the surviving children and the representatives of the deceased ones. In Cook v. Catlin, 25 Conn., 387, a testator having only nephews and nieces made a bequest “to my heirs.” The counsel who contended for a per capita distribution insisted upon the English rule, that the word heirs was descriptio personarum, and that the will must be construed as if the testator had named eaclr heir individually ; but this rule was rejected by the court on account of the different construction given to our statute of distributions in the early case of Kennedy v. Kennedy, decided in 1786, and the court held that the nephews and nieces should take per stirpes, and' not per capita. In Lord v. Moore, 20 Conn., 122, the testator had a wife and four children, to each of *476whom by name he first made specific bequests, and then gave the residue of his estate to trustees to “divide the same equally between my said wife and said children and their heirs;” and the court held that these words, in the connection in which they were used, gave to the wife a share equal only to one of the children. In Gold v. Judson, 21 Conn., 616, the testator used the following language: — “ I give to the heirs of my brother A, deceased, the heirs of my sister B, the heirs of my brother (7, the heirs of my sister E, and the heirs of my sister E, the residue of my estate, to be equally divided between said heirs, each individual alluded to having an equal portion of the same.” The intention of the testator was considered too plain on this part of the will for any doubt, and Ellsworth, J., in giving the opinion, said: — “We need not remark that the devisees are to take per capita, this being the express language of the will.'”
For the foregoing reasons we advise that the estate in question be divided into-five equal shares, and that the two surviving sisters take one share each, and that the grandchildren of the deceased sisters and brother take per stirpes.
In this opinion the other judges concurred.