Ingestive Classics
A Tribute to James C. Smith in Celebration of His 90th Birthday – Trace Conditioning in Taste Aversion Learning

Trace conditioning with x-rays as an aversive stimulus.
Psychonomic Science 9; 11-12, 1967.

Comments by Alan C. Spector (March, 2019)

My mentor, academic father, and friend, Dr. James C. Smith, celebrated his 90th birthday on the 27th of December, 2018. I wanted to take this opportunity to honor his contributions to the field of ingestive behavior by choosing one of his articles as an Ingestive Classic. Dr. Smith’s professional life was scientifically eclectic, ranging from x-ray detection in a wide variety of species, to the sense of smell in birds with their olfactory nerves cut, to detailed analyses of spontaneous feeding and drinking patterns in rats, to the aversive effects of magnetic fields in rodents, and to the functional properties of taste aversion learning with an emphasis on ionizing radiation as the unconditioned stimulus. It is this latter topic in his diverse research portfolio from which I chose a classic paper entitled “Trace conditioning with X-rays as the aversive stimulus” (4), which he published in 1967 with his doctoral student at the time, David Roll.

As historical context, in 1955, Garcia, Kimeldorf and Koelling published a paper in Science demonstrating that rats that consumed saccharin during exposure to gamma radiation subsequently avoided ingesting the sweetener when it was presented along with water (1). The avoidance lasted for weeks. For the following twenty years, researchers extensively studied the nature of this basic phenomenon, and argued about whether it was even a form of learning. Of course, it is now well documented that what Garcia et al. observed was indeed, as they initially believed, a form of associative learning that is commonly referred to as conditioned taste aversion (CTA). The taste is considered the conditioned stimulus (CS) and the agent causing visceral malaise is considered the unconditioned stimulus (US; see 2).

Decades of research have revealed that CTA has three unique functional properties in terms of associative learning. First, CTA is one of the most efficient forms of learning known. It often requires only one episode of a CS-US (i.e., taste-visceral malaise) pairing. This is thought to protect the animal from poisoning on a second occasion. In other words, it is a psychobiological manifestation of “fool me once, shame on you; fool me twice, shame on me”. After all, the animal might not be so fortunate as to survive the second time.

Second, the more novel the CS, the more effective the conditioning. This is thought to prevent animals from learning to avoid familiar foods that have been already vetted by prior experience and determined safe. Thus, if an animal eats both a familiar and a novel food close in time followed by a bout of nausea, the latter food will be the target of the animal’s avoidance on future occasions.

The third unique property of taste aversion learning is that there can be a significant delay between the ingestion of the CS and the onset of the visceral symptoms of the US. This is referred to as trace learning. The adaptive significance of this feature of CTA is that some ingested toxins may not have immediate visceral effects. The Smith and Roll (1967) paper was groundbreaking because it showed that this CS-US delay could be as long as 12 hours!

In this experiment, different groups of water-deprived animals were trained to drink water for 20 min each day. On the conditioning day, the water was replaced by a preferred concentration (0.1%) of a novel saccharin solution (CS) followed by a 100 R exposure (in 3.3 min) to X-rays (US). The different groups received the US at different time periods (0.5, 1, 2, 3, 6, 12 hours) following the end of the CS presentation. Twenty-four hours after their X-ray exposure, the animals received a 48-hour two-bottle preference test between water and saccharin, with the bottle positions switched after 24 h to control for position preferences. Control groups were treated identically except that they received sham exposure to the X-rays. Figure 1 tells the story. It is clear that when saccharin was the CS, a CS-US delay of up to 12 hours still produced some degree of CS avoidance in the 48-hour two-bottle test compared with the sham-exposed control group (the p-value at 12 hours is marginal).

Interestingly, the authors conducted a second experiment using 4% sucrose as the CS and found that the maximum CS-US interval that could lead to a significant conditioned taste avoidance in the two-bottle test was 6 rather than 12 hours. At the time, the reason for the difference was unclear, and I believe that still is the case. If I may speculate, I suspect that it may have to do with the fact that saccharin is a taste stimulus with a bitter component whose strength is concentration-dependent. It could be that at the 0.1% concentration, the saccharin generates sufficient bitterness that it becomes a more salient CS for avoidance learning.

When reading the article, please keep in mind that this experiment was conducted during a time when CTA research was still in its relative infancy, or at least childhood. As I mentioned above, it was not completely clear what the underlying mechanisms were, and this was especially true for ionizing radiation as the US (3). In this regard, Dr. Smith was a pioneer. The systematic and logical reasoning, given the prior literature, is evident in the opening paragraphs. This paper was a seminal contribution to the CTA literature. To my knowledge, this study was the first demonstration of effective taste avoidance learning with CS-US intervals of up to 12 hours. As an interesting footnote, this paper was published in a journal that, by today’s standards, would be considered far from “high impact”. Yet, of all of Dr. Smith’s publications, including those appearing in Science, he has said that it is this article that generated by far the greatest number of reprint requests (back in the days before the emergence of personal computers and electronic journals, scientists would send reprint request cards to authors via postal mail). It has been over 50 years since the publication of this article, and it is still cited today as the classic example of trace conditioning. Time is the ultimate test of good science, not the impact factor of a journal.

Happy birthday Dr. Smith and thank you for a lifetime of outstanding science.


  1. Garcia J, Kimeldorf DJ, and Koelling RA. Conditioned aversion to saccharin resulting from exposure to gamma radiation. Science 122: 157-158, 1955.

  2. Schier LA, and Spector AC. The functional and neurobiological properties of bad taste. Physiol Rev 99: 605-663, 2019.

  3. Smith JC. Radiation: its detection and its effects on taste preferences. Progress in Physiological Psychology 4: 53-118, 1971.

  4. Smith JC, and Roll DL. Trace conditioning with x-rays as an aversive stimulus. Psychon Sci 9: 11-12, 1967.