J. F. Plumel, a resident of the city and county of San Francisco, died on July 11, 1905, leaving an estate ponsisting of his separate property. His sole heirs at law were three sisters, the appellants here, and his widow, Annie Plumel, the respondent.

Two instruments, written respectively on the obverse and reverse sides of a single sheet of paper, were offered for probate as the will and codicil of the decedent, and both were admitted to probate. From the order admitting the alleged will to probate, the sisters prosecute this appeal.

The will was dated January 12, 1904. By its terms the residue of the estate, after certain bequests to the sisters and others, was given to the respondent, who is named as executrix. This instrument was not attested. It was entirely written, dated, and signed by the hand of the decedent, with the exception of the figures “190” in the date 1904. The figures “190” were printed.

Upon the back of the same sheet of paper the codicil was written. It complied with the requirements of the law regarding holographic wills, being entirely written, dated, and signed by the hand of the testator, and read as follows:

The will of January 12, 1904, being unattested and being invalid as a holographic will, because not entirely written, dated, and signed by the hand of the testator, was not, standing alone, entitled to be admitted to probate. (Civ. Code, sec. 1277 ; Estate of Billings, 64 Cal. 427 , [ 1 Pac. 701 ].) But the contention of the respondent is that the codicil incorporated in itself the terms of the will, and that, on proof of the due execution of the codicil, both documents were properly admitted to probate as the testamentary act of the decedent. This position is not, as the appellants assume, based on section 1287 of the Civil Code, which provides that “the execution of a codicil, referring to a previous will, has the effect to republish the will, as modified by the codicil. ” It is not necessary here to consider whether this section has any application to the facts of this case. Apart from its terms, it has long been settled that a will or .codicil executed in accordance with the requirements of statute may, by an appropriate reference, incorporate within itself a document or paper not so executed. (30 Am. & Eng. Ency. of Law, 2d ed. 578 ; Habergham v. Vincent, 2 Ves. Jr. 204, 228 ; Smart v. Prujean, 6 Ves. Jr. 560 ; Allen v. Maddock, 11 Moore P. C. 427 ; Brown v. Clark, 77 N. V. 369.) In Estate of Willey, 128 Cal. 1 , [ 60 Pac. 471 ], this court quoted with approval the following language from Bedfield on the Law of Wills (vol. 1, p. *264) : “The cases already referred to show very clearly that a will required to be witnessed by two or more persons, or executed with any other prescribed formalities, may, nevertheless, adopt an existing paper by reference. . . . This ‘incorporation’ of the paper referred to into the will so makes it a part of the instru meat that no distinct proof of the paper is required, or even filing, in the probate court. The proof of the will sets up and establishes the paper, as a portion of itself, by proof of the reference to the consequent incorporation.” And the court went on to say: “The principle is also substantially declared in the case of In re Soher, 78 Cal. 481 , [ 21 Pac. 8 ]. Of course, the reference must be certain, and to an instrument then in existence.”

It is no doubt true, as is stated in the Willey case, that in order to make out a case for the application of the doctrine of incorporation by reference, the paper referred to must not only be in existence at the time of the execution of the attested or properly executed paper, but that it must be referred to in the latter as an existent paper, so as to be capable of identification. But we think that in the present case there was a sufficient reference in the codicil to identify the will upon the obverse of the same sheet of paper as the instrument referred to. The later paper is designated by the testator as a “codicil,” a term which in itself implies that it is an addition to or modification of some existing testamentary paper. “A codicil is some addition to or qualification of a last will and testament. A codicil is part of a will to which it is attached or referred, and both must be taken and construed together as one instrument.” (Proctor v. Clarke, 3 Redf. 445, 448 .) By its very definition, the word “codicil” imports a reference to some prior paper as a will. And the fact that the codicil is written upon a sheet of paper containing a writing which purports to be testamentary in character is sufficient to justify the inference that such writing is the will referred to by the codicil.

In Jarman on Wills (p. *153) it is said: “It seems to have been considered, in the ecclesiastical courts at least, that the fact of the codicil being written on the same piece of paper as the prior will (though it does not in terms refer to such will), sufficiently indicates an intention to treat that as the subsisting will.” In Hobart v. Hobart, 154 Ill. 610 , [ 45 Am. St. Rep. 151 , 39 N. E. 581 ], the court said: “Where the codicil is written on the same piece of paper as the will, or clearly and unmistakably refers to the will, so as to preclude all doubt of its identity, proof of the codicil establishes the will without further proof, except such portions thereof as are revoked or altered by the codicil.” In Goods of Heathcote, 6 P. D. 30, an invalid will was offered, together with a codicil upon the same sheet of paper, beginning with the words: “This is a codicil to the last will and testament of me.” There was no further or other reference to any will, and there was no later will than the one contained upon the same sheet of paper as the codicil. It was held that there was a sufficient reference to allow the will to be probated. In view of the meaning of the word “codicil,” the language of the testatrix in the case just cited has no greater force as a reference to a prior will than has the single word “codicil” in the case at bar. In Harvey v. Chouteau, 14 Mo. 588 , [ 55 Am. Dec. 120 ], where the paper incorporated was particularly described in the codicil, but not attached to it, the court said: “If the codicil had been attached by a wafer ... I presume there would have been no room for doubt. A list of decisions -for more than one hundred and thirty years sustains this point. What is the difference between this wafer annexation of a codicil, which may not mention the previous will, otherwise than by reciting that ‘this is my codicil to my last will,’ and the case before us?”'

We are referred to no case holding that the fact that a codicil is physically attached to, or is written upon, a paper containing a prior attempted testamentary disposition, is not to be considered as tending to establish that this prior writing is the one referred to in the codicil. It is urged that the reference must be certain so as to enable the identification of the paper sought to be incorporated without the aid of evidence outside of the codicil itself, and that parol evidence is required to show that the will and codicil are written on the same sheet of paper. It is true that there is some language in Estate of Young, 123 Cal. 337 , [ 55 Pac. 1011 ], which, taken literally, tends to sustain the view that the identification of the paper to be incorporated must be possible from the face of the will or codicil alone. In that case the reference in the will was merely to “two deeds,” which were not described in any way. Any two deeds, whenever executed, would have satisfied the reference. It was held, and beyond question rightly, that parol evidence was not admissible to show that the testatrix had two particular deeds in mind. But it is to be remembered that no reference, however explicit on its face, can identify a separate paper without the production of evidence to show that the particular paper offered does correspond to the descriptive particulars named in the will. We think the correct rule was stated by this court in In re Shillaber, 78 Cal. 144 , [ 5 Am. St. Rep. 433 , 15 Pac. 453 ], quoting from Allen v. Maddock, 11 Moore P. C. 427: “A reference in a will may be in such terms as to exclude oral testimony, as where it is to papers not yet written, or where the description is so vague as to be incapable of being applied to any instrument in particular; but the authorities seem clearly to establish that, where there is a reference to any written document, described as then existing in such terms that it is capable of being ascertained, parol evidence is admissible to ascertain it, and the only question is, whether the evidence is sufficient for the purpose. Similarly, in Brown v. Clark, 77 N. Y. 369 , the court of appeals of New York said: “It is established by a long line of authorities that any testamentary document in existence at the execution of a will may, by reference, be incorporated into and become a part of the will, provided the reference in the will is distinct and clearly identifies, or renders capable of identification, by the aid of extrinsic proof, the document to which reference is made.” (See, also, Dickenson v. Stidolph, 11 C. B. (N. S.) 341.)

As we have said, the use of the word “codicil” imports a reference to some already existing document, regarded by the testator as his will. To identify that instrument, and to interpret the reference as applying to it, the surrounding circumstances may be shown. We think there is nothing in the decision in Estate of Young, 123 Cal. 337 , [ 55 Pac. 1011 ], which conflicts with the conclusion drawn from the authorities above cited; i. e. that the facts—1. That the codicil is written upon the same piece of paper as a writing purporting to be a will; and 2. That no other will is produced—may be considered as tending to identify the purported will as the one referred to.

The order appealed from is affirmed.

Shaw, J., and Angellotti, J., concurred.