The sole issue of this appeal is whether the Trial Court erred in finding that Bertha Mitchell was not entitled to worker’s compensation benefits for a knee injury. The Trial Court found that Ms. Mitchell did not prove that her knee injury was caused by a work-related accident.

On June 21, 1978 Bertha Mitchell, a custodial employee of the Orleans Parish School Board, was injured during work when she fell while picking up a mop on a wet floor. Mitchell filed suit in December of 1978 for worker’s compensation benefits, stating in her petition that the fall caused injury to her neck and back. After trial, the Trial Judge found Mitchell partially disabled, and awarded her worker’s compensation benefits plus medical expenses for her back injuries.

In 1982 Mitchell asked the School Board to pay for diagnostic treatment of her left knee. When the School Board refused to pay, Mitchell filed a second suit. The judge, who had also decided Mitchell’s first case, denied her second claim. He issued written reasons for judgment stating that: 1) had Mitchell tom her knee cartilage in her June 1978 fall, she would have complained at that time of pain in her knee but did not, and 2) on her return to work in June of 1979 she was not able to work because of her back and at that date there was no evidence of knee trauma. The Trial Judge concluded from these facts that Mitchell failed to connect her knee complaints with her June 1978 accident.

Mitchell appeals that judgment, claiming that her knee pain resulted from her back injury and alternatively, that the School Board should pay for a diagnostic test to determine whether her knee pain is related to her back injury. Mitchell also seeks penalties and attorney’s fees under La.R.S. 23:1201.2, claiming the School Board’s refusal to pay for the diagnostic test was arbitrary, capricious or without probable cause.

Causation in compensation cases is a question of fact to be found by the court on the basis of all the credible evidence. Haughton v. Fireman’s Fund American Insurance Cos., 355 So.2d 927 (La. 1978). A reviewing court may not disturb a Trial Judge’s evaluations of credibility and factual determinations unless the record reveals that the decision by the Trial Judge is manifestly erroneous. In a worker’s compensation case an employee carries the burden of establishing by a preponderance of the evidence the causal connection between any injury claimed and the work-related accident. Gremillion v. Babst Services, Inc., 418 So.2d 637 (La. 1982). The employee’s injury is presumed to have resulted from the accident, however, when the employee is able to prove that before the accident he was in good health, but *1274commencing with the accident the symptoms of the disabling condition appear and continuously manifest themselves after-wards, provided the medical evidence establishes a reasonable possibility of such a connection. Berry v. Livingston Roofing Co., 403. So.2d 1247 (La. 1981).

We do not have the benefit of the transcript of the first trial because it was lost. Accordingly, our review of the facts consists of exhibits introduced at the first trial, the Trial Judge’s handwritten notes and the transcript as well as exhibits of the second trial.

The record indicates on the day Mitchell fell in June of 1978 Dr. Pierre Espenan examined her and found that she had pain in her right lower back, right hip and right leg. She had a lumbosacral strain on an arthritic spine. He treated her for approximately two months. During this treatment he referred her to Dr. George Berkett, an orthopedic consultant, because she still had pain which Dr. Espenan attributed not to the fall but to arthritis. In a report to Dr. Espenan, Dr. Berkett reported that Mitchell’s back had improved considerably but that her neck, particularly on the right side, had not. Dr. Berkett noted that she had no history of prior injuries, illnesses or operations except for an appendectomy at age seventeen.

Dr. John J. Watermeier, an orthopedic surgeon, testified that when he began treating Mitchell in August of 1978 she had complaints of back and leg pain but no complaints of knee pain. Dr. Watermeier testified that the leg pain was on the left side which he initially thought was caused from a herniated disc condition in her back. Over a period of time, however, the back condition improved and Dr. Watermeier advised Mitchell to return to work. Mitchell returned to work in June of 1979 for approximately eight days. Dr. Watermeier’s June 8, 1979 notes state that he did not believe Mitchell, an elderly lady with severe arthritis, should be walking up and down stairs to clean the school. The first time Mitchell complained to Dr. Watermeier about her left knee was on June 27, 1979, and since that date, he testified that she has consistently complained about her left knee.

Dr. Watermeier testified that Mitchell should have a diagnostic test to determine whether her knee injury is an arthritic condition or a meniscus tear caused by her back injury. When someone sustains even a very small tear in the knee, Dr. Water-meier testified, more likely than not the knee would swell. Although he testified that in his opinion Mitchell has a meniscus tear in her knee, he also testified that he did not notice any swelling in the left knee for the first year of treatment. Nor did he notice any swelling of the left knee when Mitchell first complained about it in June of 1979. Dr. Watermeier could not say whether the tear was caused by her fall in 1978 or return to work in 1979. Because of her weak back, leg or knee from the 1978 accident, Mitchell fell twice in 1982 according to Dr. Watermeier’s notes. Her return to work and these falls, Dr. Watermeier testified, aggravated her knee condition.

At the request of the School Board, Dr. Gordon Nutik, a specialist in orthopedics, examined Mitchell in February of 1985. He also reviewed all of her medical records. He testified that Mitchell’s left knee pain was not related to the fall because more than one year occurred between the accident and Mitchell’s first complaints about her left knee. Dr. Nutik also testified that the causal connection between the accident and her knee injury was further weakened by the fact that approximately another year (August 21, 1979 to August 27, 1980) elapsed between Mitchell’s first and second complaint about her left knee. Dr. Nutik corroborated Dr. Watermeier’s testimony that if Mitchell had suffered a minor meniscal tear in her left knee at the time of her fall, she would have had on that date an irritation within that knee joint which would have caused swelling and disability of that joint. Dr. Nutik agreed with Dr. Watermeier’s testimony that the diagnostic test would determine what type of damage occurred, but Dr. Nutik further testified that this test would not reveal when the *1275damage occurred. Dr. Nutik testified that in his opinion Mitchell had degenerative arthritis of the left knee.

Mitchell testified that when she fell the pain was so severe in her neck and back that she was not concerned about her left knee. After she worked in June of 1979, she experienced pain in her back, neck and leg. Mitchell testified that her knee has become progressively worse since that date.

In view of all the evidence in the record we cannot say the Trial Judge’s evaluations of credibility and factual finding that Mitchell did not prove the 1978 fall caused her left knee injury are manifestly erroneous. The presumption of causation is not applicable because Mitchell did not prove that commencing with the 1978 fall the symptoms of a tear in her left knee appeared and continuously manifested themselves afterwards. Rather, the uncon-troverted evidence was that Mitchell neither complained nor showed symptoms of injury to her left knee until more than one year after her June 1978 fall. This medical evidence does not establish a reasonable possibility of a connection between the 1978 fall and her left knee complaints. For these reasons, the judgment of the Trial Court is affirmed.

AFFIRMED.