delivered the opinion of the court.
The judgment appealed from herein rests upon a claim of breach of contract implied from the relationship of passeng-er and common carrier that existed between plaintiff and the defendant Company. The verdict was for $10,000, and after remittitur of $3,000 judgment for $7,000 was entered for plaintiff.
Plaintiff contended that while she was attempting to alight from the car on which she was a passenger, it suddenly started causing her to fall and be injured. Defendant contended that she stepped off the car while it was in motion. Both the manner of the accident and the extent of her resultant injuries were sharply controverted. On both issues defendant had the lesser number of witnesses, and complains that the jury were erroneously instructed as to the weight to be given their testimony.
The instructions were oral and that part relating to the preponderance of evidence was as follows:
*76“The court instructs the jury that the preponderance of the evidence in the case is not alone determined by the number of witnesses testifying to a particular fact or set of facts, but the element of numbers should be considered with all the other'elements hereinafter stated; for it is in the judgment of the jury what a particular element is worth, and the evidence of the smaller number cannot be taken by the jury in ■preference to that of a larger number unless the jury can say wider their oaths that it is more reasonable, more truthful, more disinterested and more creditable (credible). •
“The court further instructs the jury in determining upon such preponderance of the evidence it should take into consideration the opportunities the several witnesses had of seeing and knowing the facts about which they testified, their conduct and demeanor while testifying, their interest or lack of interest, if any, in the result of the suit, the probability or improbability of the truth of their several statements, and, in fact, all the other- evidence, facts and circumstances proved on the trial, and from all these circumstances determine on which side is the weight or preponderance of the evidence.”
The error assigned is to the italicised part' of the instruction. In support of it appellee cites Gage v. Eddy, 179 Ill. 492, where an instruction containing the same language was held not to be misleading. In so holding, however, the court specifically referred to the fact that the complaining party had the greater number of witnesses. Here the situation was reversed, and it is manifest that the instruction could be prejudicial only to the party having the lesser number of witnesses. It must, therefore, be considered with reference to that situation to which the Gage decision cannot be held to apply; for it does not reasonably follow that because in one case an instruction has no tendency to mislead the jury or injure the complaining party, it will not have such effect when applied to *77another situation or an entirely different state of facts.
In the present case, we think, giving the italicised part of the instruction constituted prejudicial error. Without it the instruction was in the approved form on the subject of preponderance of evidence, telling the jury the various elements they may consider in determining where it lies, including that of the number of witnesses testifying to a particular fact of set of facts, and that that element could not be considered alone. But the italicised part is inconsistent therewith. It tells the jury, in effect, that applying a certain test they may after all decide the case on that element alone. The test is that they cannot take the testimony of the smaller number unless it possesses all of four characteristics; in other words, if any of the four is lacking, then the jury must find according to the greater number of witnesses, for the instruction tells them, in effect, that in such a case they cannot take the testimony of the smaller number.
In this form the instruction violates several fundamental principles. It invades the province of the jury. It conflicts with the doctrine of the burden of proof. It practically directs a verdict upon incomplete elements of the case, and it requires the application of an incorrect test.
As to said test, it is not true that before a jury can accept the testimony of the smaller number of witnesses they must find it is the more “disinterested.” Applying that test, the jury may have concluded that because one of the defendant’s witnesses to the accident was its conductor, and another, called as an expert on the question of injuries, had testified for it off and on in that capacity for nearly twenty years, it could not find their testimony was the more “disinterested,” and, therefore, could not decide for defendant on either of the main issues on which they testified. It frequently happens that interested wit*78nesses had a better opportunity to know the facts about which they testified; that as to physical facts they were in a better position to hear or see, and for determining a scientific fact they had superior education, greater intelligence and larger esperience. But under the test furnished in this instruction, even though their testimony appeared more reasonable, truthful and credible than that of a greater number testifying differently, yet the jury could not accept it in preference to the latter unless they could’ also say it was more “disinterested.”
Nor is this objection cured by the fact that elsewhere in the instruction the jury were told they should consider the element of numbers with other elements enumerated therein. The italicised part of the instruction is so phrased that they may decide the case upon that alone.
Besides, in its peremptory form, it encroaches on the province of the jury to determine for themselves the weight that should be given to the testimony, a privilege the court should not interfere with (Chicago & A. R. Co. v. Robinson, 106 Ill. 142), and with respect to which no artificial rule can be laid down. Clevenger v. Curry, 81 Ill. 432. We think it was calculated to impress the jury with the idea that in this case the element of numbers was of superior and controlling importance.
But it also conflicts with the rule as to the burden of proof. The jury were properly told that if the weight of the evidence was evenly balanced between plaintiff and defendant on the material issues, plaintiff could not recover and the jury should find for defendant. But that part of the instruction under consideration told the jury that they could not find for defendant unless they found the testimony of its witnesses was more truthful, more credible, etc., than that of plaintiff’s, whereas, if their testimony was of *79equal weight, defendant, under the rule of the burden of proof, was entitled to the verdict.
But this was not the only prejudicial error in the case. Two physicians, called as experts by plaintiff on the question of damages, were permitted to express opinions manifestly based upon subjective conditions. One of the alleged injuries from the accident was its effect upon plaintiff’s mind, particularly her power of memory. One of said experts testified that from his examination he found ‘ ‘ there was a lack of mental and physical confidence, a lack of memory.” But it is evident that he based such conclusion on plaintiff’s conversation. Testifying he said: “The woman in her conversation showed,” and was interrupted by objection from defendant’s counsel on the ground that she could control her conversation. Plaintiff’s counsel then suggested a rule of guidance in such a matter in the following language: “Supposing a person is telling the truth in their conversation, that is an objective symptom. Doctor, you are to limit this entirely to objectiye symptoms.” The court evidently adopting this suggestion as a correct criterion, overruled the objection, and thereupon the doctor answered as above.
The other physician, while testifying as to her condition found by his examination, was, over defendant’s objection, permitted to say: “While in conversation with Mrs. Newcomb, it was evident that she had not regained her normal condition as I previously knew her.” He was then asked to state her mental condition at that time and said that, compared with the normal individual, he would say that she was “twenty-five per cent, minus.” The motion to strike out this answer was overruled.
It is apparent that the examination relied upon by each physician was made with a view to testifying as an expert, and it is settled law that an expert witness called under such circumstances must base his opinion upon objective and not subjective conditions, and that *80self-serving statements made by the injured party at such a time are incompetent. Greinke v. Chicago City Ry. Co., 234 Ill. 564; Shaughnessy v. Holt, 236 Ill. 485. What statements she made in such conversations were not given to the jury, but they might just as well have been if the experts were permitted to draw their inferences therefrom. One’s conversation is usually under his control, and it was in the power of appellee to make statements from which alone the physicians might draw such inference. We think this evidence is not only inadmissible but prejudicial, for as said in the Shaughnessy case, supra: “Appellee was allowed to enhance her damages by proof of her own incompetent, self-serving statements, to what extent it is impossible to judge.” On the question of damages much stress was laid upon appellee’s mental condition. The verdict was so large that her counsel found it advisable to enter a remittitur, but, in view of the stress laid in the case upon this one element of damages and the importance the jury might give to the testimony of experts on the subject, we are not prepared to say that the remittitur cured the error. Lyons v. Chicago City Ry. Co., 258 Ill. 75.
Other errors are assigned and argued, but we need not consider them as they are not likely to arise in another trial. Upon those considered, we think the judgment should be reversed and the cause remanded for a new trial.
Reversed, and remanded.