To those familiar with the film exploits of Lassie, Rin Tin Tin or Benji, it might seem as if there's no end to what a dog can be trained to do. Such seemed to be the case with a remarkable Dalmatian named Sunny the Wonder Dog, which has become something of a media celebrity because of his reputed ability to add, subtract, multiply and divide.

In public appearances on television, radio and in schools and libraries with his master, Sunny has astounded his audiences by barking out correct answers to problems written on flash cards By his master, Jim Todd, a retired engineer in San Jose.

Furthermore, Sunny seemed able to calculate square roots, cube roots and other complex equations. The dog also seemed able to respond to math questions posed in Spanish, Portuguese and Yiddish.

Some of our skeptical colleagues in the San Francisco area decided it would be fun to test Sunny's mathematical skills, and Todd happily agreed.

He and Sunny showed up for the test looking both relaxed and confident. The first hint of difficulties came just before the actual test was about to begin.

Todd stated that Sunny was having difficulty that day in barking out the numbers one, two, nine and 10. (Normally, Sunny will bark out the answer to practically any problem so long as it is between one and 10.) Hence, on this day Sunny could only be presented with problems that had answers between three and eight.

The mutually agreed agenda was that for the first 15 minutes Todd could demonstrate Sunny's skills without interference; then for the next 15 minutes Bay Area skeptic Don Henvick would attempt to test Sunny's knowledge.

The easygoing Todd put on an impressive show of Sunny's skills. Holding up cards with a number written on each, he would ask, "How much is three and four?"

Sunny would bark out the correct amount. Problems like this were repeated several times, usually with Sunny getting the right answer. Occasionally, the dog got the answer wrong, sometimes by barking hesitantly or quietly, or with unclear "enunciation." When this happened, Todd would berate him and Sunny would usually get the correct answer the second time.

It was a stunning, if not altogether convincing, performance. There remained, however, an alternative explanation--that Sunny might have been receiving cues from Todd.

There was a very simple way to test this possibility: Have someone other than Todd present the cards to Sunny, exactly as Todd does, but present them in such a way that only the dog and a video camera could see what was written on the card. In addition, Todd would stand behind the dog. This would eliminate the possibility of any unconscious cuing to the dog.

Once this was done, Sunny's math abilities deteriorated immediately.

The first card was shown to the dog and Sunny was asked to name the number on the card. Sunny barked eight times. The card had a five on it. Sunny barked 11 times for the next card; it was a three.

The positions then were rearranged, so that Sunny was in between Todd and Henvick, who was presenting the cards to the dog. A movable blackboard screened Todd's face from Sunny.

Strangely, Sunny paid no attention to Henvick and the cards; he kept turning to face Todd. The dog's mathematical skills were now failing badly: seven plus three made nine; four plus five made 13.

It soon became evident that Sunny was simply barking randomly, unsure of what it was he was supposed to do. An attempt at this point to answer math questions posed in Spanish was likewise a failure. "* Dos y dos* " (two plus two) elicited seven barks.

Todd was now becoming visibly upset. He began explaining how Sunny was getting tired. His sincerity was painfully obvious; he really expected Sunny to be able to answer correctly.

Exactly how Sunny was able to perform so well in the first place was never precisely determined. But it seemed likely that Sunny had been trained to begin barking whenever Todd held up a card and began talking to him.

After watching the videotape of the test, it became apparent that while accumulating the required number of barks, Todd stood very stiffly; his breathing became exaggerated and was clearly audible.

As soon as the required number of barks were received, Todd broke his stance and lowered his breathing. In addition, he brought the card down, an obvious sign to the dog to stop.

Although we think Sunny cannot correctly answer questions except when Todd cues the answer, it is fun to wonder just how much dogs can understand from the subtle clues from their masters.