The sole issue presented by this appeal is whether caveator presented a prima facie case that testator’s will was the product of undue influence in order to survive propounders’ motions for a directed verdict. The Court of Appeals held that caveator’s evidence of undue influence was insufficient to take the case to the jury. We reverse.
To constitute undue influence within the meaning of the law, there must be more than mere influence or persuasion because a person can be influenced to perform an act that is nevertheless his voluntary action. In re Will of Frank, 231 N.C. 252, 56 S.E. 2d 668 (1949), rehearing denied, 231 N.C. 736, 57 S.E. 2d 315 (1950). For the influence to be undue,
*54“ ‘there must be something operating upon the mind of the person whose act is called in judgment, of sufficient controlling effect to destroy free agency and to render the instrument, brought in question, not properly an expression of the wishes of the maker, but rather the expression of the will of another. It is the substitution of the mind of the person exercising the influence for the mind of the testator, causing him to make a will which he otherwise would not have made.’ ” In re Will of Kemp, 234 N.C. 495, 498, 67 S.E. 2d 672, 674 (1951), quoting In re Will of Turnage, 208 N.C. 130, 131, 179 S.E. 332, 333 (1935); see generally, Wiggins, Wills and Administration of Estates in North Carolina § 55 (1964).
In a caveat proceeding, the burden of proof is upon the pro-pounders to prove that the instruments in question were executed with the proper formalities required by law. In re Will of West, 227 N.C. 204, 41 S.E. 2d 838 (1947). Once this has been established, the burden shifts to the caveator to show by the greater weight of the evidence that the execution of the will was procured by undue influence. Id.
The battle by one person to overpower and overcome the free will, agency, wishes and voluntary action of another is very often difficult to prove because, after testator’s death, only circumstantial evidence remains from which the trier of fact must decide whether the battle in fact occurred and whether testator was on the losing side. Caveator must rely on inferences from the surrounding facts and circumstances that arise on the evidence in his effort to prove that undue influence existed at the time testator executed his last will and testament thereby causing him to execute a will that he otherwise would not have executed. The more adroit and cunning the person exercising the influence, the more difficult it is to detect the badges of undue influence and to prove that it existed. In re Will of Beale, 202 N.C. 618, 163 S.E. 684 (1932).
It is impossible to set forth all the various combinations of facts and circumstances that are sufficient to make out a case of undue influence because the possibilities are as limitless as the imagination of the adroit and the cunning. The very nature of undue influence makes it impossible for the law to lay down tests to *55determine its existence with mathematical certainty. In re Will of Beale, supra.
Several of the factors that are relevant on the issue of undue influence include:
“1. Old age and physical and mental weakness.
2. That the person signing the paper is in the home of the beneficiary and subject to his constant association and supervision.
3. That others have little or no opportunity to see him.
4. That the will is different from and revokes a prior will.
5. That it is made in favor of one with whom there are no ties of blood.
6. That it disinherits the natural objects of his bounty.
7. That the beneficiary has procured its execution.” In re Will of Mueller, 170 N.C. 28, 30, 86 S.E. 719, 720 (1915).
Of course, no matter how difficult the task may be, the burden of proving undue influence is on the caveator and he must present sufficient evidence to make out a prima facie case in order to take the case to the jury. The test for determining the sufficiency of the evidence of undue influence is usually stated as follows: “[i]t is ‘generally proved by a number of facts, each one of which standing alone may have little weight, but taken collectively may satisfy a rational mind of its existence.’ ” Id. at 29, 86 S.E. at 719, quoting In re Will of Everett, 153 N.C. 83, 87, 68 S.E. 924, 925 (1910).
For example, proof that on the evening of the day on which the will was executed the testator came home drunk, and that in the will he left his property to some of his nephews and a niece, to the exclusion of other nephews and nieces, are circumstances to be considered on the issue of undue influence, but standing alone they are not enough to take the case to the jury. In re Will of Harris, 218 N.C. 459, 11 S.E. 2d 310 (1940). Proof that testator was old, suffered from chronic ailments, had a poor memory and used narcotics are circumstances to be considered. However, all of these factors relate to the existence of a poor physical and mental condition which would make testator susceptible to undue in*56fluence, but standing alone these factors are insufficient to make out a case of undue influence. In re Will of Ball, 225 N.C. 91, 33 S.E. 2d 619 (1945). Proof that testator stated four years after the execution of the will that he had allowed others to take advantage of him and lead him to make the will that was ultimately brought into question is a circumstance to be considered, but standing alone it is insufficient to make out a case of undue influence. In re Will of Turnage, supra.
Therefore, the most that can be said is that a prima facie case of undue influence consists of evidence by caveator of that combination of facts, circumstances and inferences from which a jury could find that the purported last will and testament is not the product of testator’s free and unconstrained act, but rather that it is the result of an overpowering influence exerted by someone on the testator sufficient to overcome testator’s free will and agency and to substitute for it the will and wishes of that other person, so that testator executed a will that he otherwise would not have executed. When such evidence exists, then the case must be submitted to the jury for its decision as to whether or not the last will and testament of the testator was in fact the product of undue influence.
Considering the evidence in the light most favorable to the caveator, In re Will of Ball, supra, and giving him the benefit of all reasonable inferences arising on the evidence, we believe caveator presented sufficient evidence of undue influence to survive propounders’ motion for a directed verdict made at the close of all the evidence in order to take the case to the jury.
Caveator’s evidence, when presented in chronological order, indicated that the testator executed a number of wills and codicils prior to the will and codicil in question. The first will was executed in 1962. In this will, testator devised one-half of his estate to Mrs. Andrews absolutely. The remainder he devised in trust with his son as beneficiary. When his son reached age twenty-eight, the trust was to terminate and his son was to receive the principal. However, if the son died prior to the termination of the trust, the principal was to be given to his son’s issue, if any; if none, then to Mrs. Andrews. If she was not living, then the principal was to go to Mrs. Andrews’ son, Michael Jad Mahaley. Mrs. Andrews was named executrix.
*57In 1965, testator executed a codicil to the 1962 will which appointed Mrs. Andrews and R. F. Hoke Pollock (Mr. Andrews’ attorney for many years) as co-executors of the will. In 1966, testator executed a second codicil to the 1962 will in which he removed Mrs. Andrews as co-executor and appointed Pollock as sole executor. The codicil further provided that if testator’s death should result from any cause other than natural causes, his wife, Mrs. Andrews, should receive nothing under the will. He also directed that an autopsy be performed to determine the cause of his death.
In 1970, testator executed a will revoking all prior wills and codicils. In this will, testator devised his estate to a trustee to pay one-half of the income to Mrs. Andrews for life and one-half to his son for life. Upon the deaths of the income beneficiaries, the principal was to be distributed to testator’s grandchildren in equal shares. Pollock was appointed as executor.
From the above pattern of dispositions, it can be seen that the positions of Mrs. Andrews and her son under testator’s various wills and codicils worsened between 1962 and 1970. Under the 1962 will, Mrs. Andrews was to receive a one-half interest in the estate in fee simple, but under the 1970 will she was to receive only an income interest in one-half of the estate for her life. Under the 1962 will, testator’s stepson was to receive the one-half of the estate which had been devised in trust to testator’s son only if testator’s son died before the trust terminated and he had no surviving issue and provided that Mrs. Andrews predeceased him. Under the 1970 will, testator’s stepson was to receive nothing.
These changes in the disposition of his estate give rise to an inference that all was not well in the relationship between testator and his wife between 1962 and 1970. There is other evidence in the above summary from which this same inference can again be drawn. Mrs. Andrews was named sole executrix in the 1962 will. In the 1965 codicil, she and Pollock were named co-executors. In the 1966 codicil, Pollock was named sole executor. In the 1970 will, Pollock was again named sole executor. Also, in the 1966 codicil, testator wanted his wife to receive nothing if he died from any cause other than natural causes and he wanted an autopsy performed to determine the cause of his death.
*58Caveator’s evidence then tended to show that many changes occurred between 1969 and 1976. From this evidence we note that caveator presented evidence on each of the seven factors listed above as relevant to the issue of undue influence.
(1) Testator was old and was physically and mentally weak during the last few years of his life. He was seventy-seven years old at the time of his death in 1976. He had had a heart attack in 1969, and after that he was afflicted with high blood pressure and diabetes. He had been obese for many years. Evidence from his barber tended to show that testator had always been quite particular about his appearance until about 1973 when he then showed less concern. Other witnesses testified that he was not his normal self in 1974 and 1975 and was not as alert as he formerly had been. On one occasion, testator apparently was not aware that he owned certain property.
(2) Testator lived with his wife and was subject to her constant association and supervision. A former employee of the testator testified that he visited the Andrews’ home in 1975. The testator was relaxed at first, but when Mrs. Andrews appeared he became quite nervous, had tears in his eyes and could not speak. A real estate dealer indicated that in 1975 and early 1976 he negotiated a purchase of some property owned by the Andrews. He testified that most of his discussions were with Mrs. Andrews and on one occasion testator was not aware that he owned certain property that the dealer desired to purchase.
(3) Other people had little or no opportunity to see the testator. The manager of the Sheraton Motor Inn in Southern Pines testified that he purchased some property from the Andrews in 1975. In negotiating the purchase of the property, he met with testator on one occasion and after that he called the testator six or seven times but could not speak with him. On each of those occasions, Mrs. Andrews said that he was resting or had had a bad night. She told him to contact her attorneys in Charlotte who would handle the closing and apparently this was done. Testator later saw the manager at the Motor Inn and asked him why he had not been called about the purchase of the property.
Another witness indicated that he tried to get in touch with testator at testator’s request, but, when he called, Mrs. Andrews would not allow him to speak with her husband. Other witnesses *59testified that they left messages to have calls returned when they could not speak with testator, but their calls were never returned.
(4) The 1974 will and 1975 codicil were different from and revoked prior wills. (5) These two instruments were made in favor of one with whom there were no ties of blood (Mrs. Andrews and her son — testator’s stepson) and (6) to the extent those instruments increased the positions of Mrs. Andrews and her son, they disinherited a natural object of testator’s bounty (his son).
In the 1974 will, testator revoked all prior wills and bequeathed all of his tangible personal property to his wife if she survived him; if not, then to his son if his son survived him, and if he did not, then to his son’s surviving issue and testator’s stepson in equal shares. All of the furniture, household goods, silverware, china and ornaments were acknowledged to be the property of his wife. Testator devised one-half of his adjusted gross estate to his wife in such a manner as to take advantage of the maximum marital deduction allowed under the relevant provisions of the federal estate tax. The rest of his estate was devised in trust for the benefit of his son. He was to receive the income for life and, at his death, the principal was to be divided equally between testator’s stepson and testator’s grandchildren. A spendthrift provision was- attached to this trust whereby the income was to be paid to Mrs. Andrews and the principal beneficiaries in the event an income beneficiary tried to sell or transfer his interest in the trust. The testator appointed his wife as executrix. In July, 1975, the testator executed a codicil to the 1974 will in which he altered the provisions of the trust to provide that upon the death of his son, Mrs. Andrews, if she were still living, was to receive the income from the trust for her lifetime.
Also, in 1975, deeds were executed transferring real property that testator owned in the Southern Pines area from testator individually to testator and his wife as tenants by the entirety. Apparently, this was done because testator had read a book entitled, “How to Avoid Probate.”
From the above pattern of dispositions it can be seen that after the positions of Mrs. Andrews and her son suffered in the codicils to the 1962 will and the 1970 will as compared to the 1962 *60will (as summarized above), there was then an improvement in their positions under the 1974 will and 1975 codicil.
Under the 1974 will, testator’s stepson, who was to receive nothing under the 1970 will, was to receive a one-fourth interest in the principal of the trust set up under that will (divided equally between the stepson and testator’s three grandchildren) which was to contain one-half of the testator’s adjusted gross estate. Also, he was to receive one-fourth of all the tangible personal property (again to be divided equally between the stepson and testator’s three grandchildren) if both testator’s son and his wife predeceased the testator.
Mrs. Andrews was, under the 1974 will, acknowledged to be the owner of all the furniture, household goods, silverware, etc. She was to receive all of the tangible personal property if she survived the testator. After these dispositions, she was to receive one-half of testator’s adjusted gross estate. Also, she was named sole executrix under this will. The significance of this appointment was stated by Judge (now Justice) Carlton in his dissenting opinion in this case in the Court of Appeals:
“The will vests the power in the executrix to determine which property shall go into which half of the.estate. In other words, Mrs. Andrews could choose that portion of the property to be allocated to her.” In re Will of Andrews, supra at 96, 256 S.E. 2d at 258 (Carlton, J. dissenting).
The 1974 will also provided that, “my wife’s share shall not be reduced by any estate, inheritance, transfer, succession, legacy or similar taxes.” Thus, under the 1974 will, Mrs. Andrews was to receive a one-half net interest in fee simple undiminished by the payment of state and federal inheritance and estate taxes. Under the 1970 will, Mrs. Andrews would have received a one-half income interest and no fee simple interest, and under the 1962 will Mrs. Andrews would have received one-half of testator’s estate after the payment of estate and inheritance taxes. From the record it appears that, at his death, testator’s estate was worth approximately $1.5 million. Considering this, the increase in value to Mrs. Andrews under the 1974 will as compared to the provisions for her under the earlier wills can readily be seen.
*61In addition, the 1975 codicil provided that if Mrs. Andrews survived testator’s son, she was to receive his income interest under the trust which contained the other one-half of testator’s adjusted gross estate. And finally, the real property that had been transferred from testator individually to testator and Mrs. Andrews as tenants by the entirety would be hers outright as the surviving tenant.
It may be that the real property was changed to ownership as tenants by the entirety and that the 1974 will giving Mrs. Andrews one-half of the adjusted gross estate were done to minimize estate and inheritance tax consequences. However, the issue is whether the changes in the will were made because testator wanted them made or because Mrs. Andrews wanted them made. If Mrs. Andrews had them made by overcoming the free will and agency of the testator thereby causing him to make a will which he otherwise would not have made, then such action constituted undue influence. As was stated in In re Will of Turnage, supra at 132, 179 S.E. at 333:
“To constitute such undue influence, it is not necessary that there should exist moral turpitude, but whatever destroys free agency and constrains the person, whose act is brought in judgment, to do what is against his or her will, and what he or she otherwise would not have done, is . . . [undue influence] in the eye of the law.”
(7) The evidence gives rise to the inference that Mrs. Andrews procured the execution of the 1974 will. The 1962 will and the codicils thereto, as well as the 1970 will, were drafted by R. F. Hoke Pollock and executed by the testator in Southern Pines. Pollock had been testator’s attorney for some time prior to the execution of the 1962 will. Mrs. Andrews was not present when these instruments were executed. The 1974 will and 1975 codicil were drafted by Paul Wyche, a Charlotte attorney, who was employed by Belk Stores Services, Inc. He had apparently been recommended to draft the will by David McConnell, who was at that time Chief Counsel of Belk Stores Services, Inc. Apparently, Mrs. Andrews had been connected in some employment situation with Belk Stores before her marriage to testator and, with respect to another transaction, the evidence reveals that Mrs. Andrews’ attorneys were located in Charlotte.
*62After talking by telephone with testator and receiving some information from McConnell on the Andrews’ estate, Wyche drafted the 1974 will (Belk’s permitted him to have private clients so long as there was no interference with his work for the company) and mailed it to the testator. At that time, Wyche knew of the 1962 will, but he did not know of the two codicils to that will and he did not know of the 1970 will. As noted above, these were the instruments in which the positions of Mrs. Andrews and her son suffered as compared to the 1962 will.
Testator telephoned Wyche and said that he would like to meet him in Charlotte to execute the will that Wyche had prepared. Testator and his wife then travelled to Charlotte where testator executed the will. Apparently, testator had never seen Wyche before that meeting in Charlotte. This manner of preparing and executing a will for a $1.5 million estate appears unusual and while these facts may cast no questions of improper action upon Mr. Wyche, they do give rise to an inference that Mrs. Andrews procured the execution of the 1974 will.
Wyche also prepared the 1975 codicil and the deeds creating the relationships of tenants by the entireties in real property that formerly testator had owned alone. These were all prepared by Wyche after consultations with testator. Wyche drafted the 1975 codicil after travelling to Pinehurst at testator’s request where he spent the day discussing testator’s estate with testator and his accountant. The Court of Appeals relied on this testimony by Wyche (a witness for the propounders) to counter the other evidence in the case that was favorable to the caveator in order to arrive at the conclusion that, “[t]he import of his [Wyche’s] testimony was that testator acted freely and knowingly in executing this will and codicil.” In re Will of Andrews, supra at 93, 256 S.E. 2d at 256.
Reliance on this testimony by the Court of Appeals is misplaced and its conclusion is erroneous. It is our function, in a case such as this, to consider all of the evidence but to consider it in the light most favorable to the caveator, deem his evidence to be true, resolve all conflicts in his favor and give him the benefit of every reasonable inference to be drawn in his favor. See, Rappaport v. Days Inn of America, Inc., 296 N.C. 382, 250 S.E. 2d 245 (1979). Even though there may be contradictions and conflicts in *63some or all of the evidence in the case, if caveator had sufficient evidence of undue influence so that a jury could (if it believed his evidence and his version of the facts) find for the caveator, then the motion for a directed verdict should be denied and the case sent to the jury so that it can resolve the disputed issue of fact. See, Rappaport v. Days Inn of America, Inc., supra.
It may be, as propounders’ evidence tended to show, that testator was in control of his own affairs until his death and that the 1974 will and the 1975 codicil were the result of testator’s own wishes to plan his estate wisely and to minimize the estate and inheritance tax consequences. It may be that the will and codicil in question were the result of Mrs. Andrews’ overpowering and undue influence to have testator execute the will that she wanted him to execute that would benefit her and her son more than the previous wills and would also minimize the tax consequences to his estate and thus, indirectly, to her. That is the issue that the jury had to resolve because, upon correct application of the test enunciated in Rappaport v. Days Inn of America, Inc., supra as set forth above, we hold that, while no single piece of evidence alone would have been sufficient, caveator presented a combination of facts from which a jury could conclude that testator’s 1974 will and 1975 codicil were procured by the undue influence of Mrs. Andrews.
Judge Hairston submitted this case to the jury and it has spoken. The verdict and judgment for the caveator shall be reinstated. The Court of Appeals is reversed.
Reversed and remanded.
Justice CARLTON took no part in the consideration or decision of this case.