The principal question presented by this appeal by writ of certiorari is whether the respondent who suffered a- non-industrial accident has proved by a preponderance of the evidence that his present condition stems from a prior industrial accident.
The respondent, Charles Hazelton, an automobile refrigeration mechanic, injured his back in September, 1969, while working underneath a car during the course of his employment with petitioner, Valley Dodge Company, which was insured by Employers Mutual Liability Insurance Company (Employers Mutual). The accident occurred on Hazelton’s second day of employment and thereafter he never returned to work for Valley Dodge. Hazelton, whose condition was diagnosed as low back sprain, was hospitalized and after discharge received additional, extensive medical treatment.
In the early part of July, 1970, respondent opened his own garage, a business he had engaged in prior to working for Valley Dodge. On July 25, while reaching down to turn off a valve something snapped and he experienced pain. Two days later he slipped on a grease slick, nearly losing his balance, and in doing so, jarred his neck and back. This near fall produced rather severe pain.
Following the September, 1969, injury, a claim for workmen’s compensation benefits was filed and it was subsequently accepted by petitioner Employers Mutual. In August, 1970, Employers Mutual terminated all benefits on the basis that Hazelton’s then current condition was not related to the industrial accident for which it was liable. Two hearings were held, resulting in an award for unscheduled permanent partial disability of five percent. The In*518dustrial Commission affirmed the award and petitioners have appealed.
Petitioners contend that it was error for the Industrial Commission to impose full liability upon them for the entire resulting disability. Petitioners’ reasoning is as follows: Hazelton sustained a new injury in July, 1970, to the same part of his back as was injured in the September, 1969, incident ; the effects of the new, non-industrial injury (for which they are not responsible), coupled with the original injury, and the chronic obesity problem, all resulted in a permanent disability which cannot be medically apportioned. In view of the fact that respondent has failed to prove which event caused his disability or what part is apportionable to the first injury, the award, according to petitioners, is based on conjecture and speculation and as such it must be set aside.
Petitioners’ contention is predicated on the findings of the hearing officer, in particular :
“4. ... On July 25 and 27, 1970 [applicant] reinjured his back in the same area as the injury of September 9, 1969
* * ‡ ‡ ‡ ^
“8. That applicant has a pre-existing condition described as chronic obesity which retarded his recovery herein, and is a contributing factor to his present physical impairment.
“9. That applicant has a 5% physical functional impairment as a combined result of the industrial injury of September 9, 1969, the two non-industrial injuries of July, 1970, and his chronic obesity.”
The first determination that we must make is whether the medical evidence supports the conclusion that the events occurring in July constituted a new and independent injury and if so, whether the evidence supports the finding that the disability cannot be medically apportioned between the two accidents.1
A painstaking examination and consideration of the medical testimony in its entirety 2 reveals the following:
Dr. Sidney Stovall, an orthopedic surgeon, testified at the first hearing that he agreed with a group consultation report which stated that Hazelton had a five percent functional loss. This report does not state, as respondent suggests, that Hazel-ton’s disability is attributable to the September 1969 injury. However, in a letter written by Dr. Stovall, which is part of the record, he stated: “I feel that this is an aggravation of his preexisting condition or you might say it is an acute exacerbation rather than a new injury.” At the time of the hearing when examined concerning this alleged aggravation, the doctor, in response to a hypothetical question containing, among other material facts, the facts of the July incidents, could not state with reasonable medical probability that the symptoms experienced at the end of July were an exacerbation of the original injury.
Considering Dr. Stovall’s testimony as a whole we fail to sec how it supports respondent’s theory that the events in July merely constituted an exacerbation or aggravation of the original injury.
Dr. John K. Kelly, a neurological .surgeon, was asked his opinion as to the percent of disability Hazelton sustained. To eliminate any confusion that the doctor related the disability to the original injury, a reproduction of some of the questions and answers reveals actually what the doctor’s opinion was:
“Q. Do you have an opinion based on reasonable medical probability whether he has any percentage of physical disability of the whole man from the injury of September 9, 1969?
“A. Yes.
“Q. What is that opinion?
“A. As to a percentage of disability ?
“Q. Right.
*519“A. It could be, I would say, 10 to 15 percent.” (Emphasis added.)
The doctor’s qualifying question (see emphasized portion of above quotation) and his corresponding answer clearly show that his opinion is devoid of any reference as to what injury caused the disability. Furthermore, when asked directly whether he had any permanent physical disability from the grease slipping incident, the doctor said:
“A. I think it would be hard for me to distinguish between disability due to an incident when he slipped in the grease and his original injury of September of 1969. I think that it is possible that the incident, when he slipped in the grease may have aggravated his original injury.”
Dr. L. L. Tuveson, an orthopedic surgeon who examined respondent only once, which was prior to July, 1970, agreed with the group consultation report.
Dr. W. A. Bishop, an orthopedic surgeon who participated in and co-authored the group consultation report, unequivocally stated that the two July incidents were a re-injury to the back or a new injury. Dr. Bishop could not separate the effects of the two injuries by stating what percent of the total disability was attributable to one or the other injuries, though he did say “it was a good possibility” that at least a part of the five percent disability was related to the September injury.
Having considered all of the pertinent medical testimony, and the Commission having the obligation to resolve the medical conflict, if any, we hold that the Commission’s findings that the July incidents constitute a new and independent injury and by implication that the disability is un-apportionable, are reasonably supported by the evidence.
We note that Hazelton’s chronic obesity which has retarded his recovery and contributed to his present physical disability is a preexisting condition and the fact that it is impossible to designate what part of the resulting disability is attributable to this condition is of no consequence. An employer takes his employee as he is and “if an injury, operating on an existing bodily condition or predisposition, produces a further injurious result, that result is caused by the injury.” Murray v. Industrial Commission, 87 Ariz. 190, 349 P.2d 627 (1960).
We therefore hold that the obesity problem is immaterial to petitioners’ apportionment contention.
We now turn to petitioner’s final contention, relating to whether respondent proved his claim. It is a well-established proposition of law that a claimant must prove all of the elements of his claim by a preponderance of the evidence. In re Estate of Bedwell, 104 Ariz. 443, 454 P.2d 985 (1969). This burden of proof requires an affirmative showing that the injury did in fact produce the resulting disability. Thomas v. Industrial Commission, 12 Ariz.App. 23, 467 P.2d 247 (1970). When the result of an accident is not readily apparent, the causal connection between the injury and the disability must be established by expert medical testimony. Bedel v. Industrial Commission, 5 Ariz.App. 470, 428 P.2d 134 (1967). Medical conclusions as to causation must be based upon probabilities rather than upon possibilities. Helmericks v. Airesearch Mfg. Co., 88 Ariz. 413, 357 P.2d 152 (1960). Thus, if there are two or more possible causes for a disability and the medical testimony adduced to establish causality is couched only in terms of possibilities, then the claimant has not met his burden of proof. See Beougher v. Industrial Commission, 12 Ariz.App. 469, 472 P.2d 80 (1970).
Since the testifying physicians could neither say to a degree of medical probability that Hazelton’s resulting disability was caused by the September, 1969, injury, rather than by the injuries occurring in July, 1970, nor could they apportion the disability with any degree of certainty, we hold that respondent has failed to carry his burden of proof, and therefore, the award of the Industrial Commission, plac*520ing the liability for the respondent’s present disability upon petitioners is patently contrary to our workmen’s compensation law.3
Award is set aside.
HAIRE, Chief Judge Division 1, and EUBANK, J., concur.. The question of the obesity problem will be dealt with following a discussion of the above questions.
. Of the five physicians who testified, only the opinions of four are relevant to the issues presented for our consideration.
. If the petitioners had been the employer and insurance carrier at the time of the second injury, then they would be liable for the resulting disability in spite of an inability to apportion. Lockhart v. Industrial Commission, 15 Ariz.App. 209, 487 P.2d 430 (1971).