Mark Rilling. The MYSTERY of the VANISHED CITATIONS

The Search for the Engram Before McConnell and the 1960s

The Missing 1960s Decade

Memory Transfer: McConnell's Blind Alley

McConnell's Struggle With Critics Over Invertebrate Learning

   Controlling for Experimenter Bias

   Controlling for Pseudoconditioning and Sensitization

McConnell's Origin Myth for Planarian Learning

Escaping Peer Review as a Celebnty-Sclentist

The Worm Runner's Digest: Peer Review Versus the 1960s Counterculture

Attracting the Unabomber. While Overselling Behavior Modification

   The Assassination Attempt

   Overpopularizing Behavior Modification

Conclusion: Some Historical Lessons for Today From the 1960s

References

Author's note. I thank Charles Abramson, Bernard Agranoff, Jeff Bitterman, Jay Best, Francis Crinella, Arni Golub, Donald Dewshury, Thomas Nelson, William McKeachie, Robert Sommer, and especially Marlys Schutjer for interviews, correspondence, reprints, and suggestions. I thank John Popplestone and Marion White McPherson for access to the McConnell letters and articles at the Archives for the History of American Psychology, University of Akron, OH. The letters cited in the text are located in the McConnell file at the archives. An earlier version of this article was presented at the International Conference on Comparative Cognition, Melbourne, Florida, March 1995.

Correspondence concerning this article should be addressed to Mark Rilling, Department of Psychology, Michigan State University, East Lansing, MI 48824. Electronic mail may be sent via Internet.

A piece of scientific research, like a work of fiction or a set of laws is never the result of a single individual's efforts. It is rather the end product of myr iad thoughts and actions and discoveries that have gone before. What debts a researcher does not owe to his contemporar ies and his immediate mentors, he owes to his predecessors. In truth as Newton observed we all stand on the shoulders of giants of the past and present. (James McConnell, 1956, preface to dissertation).

McConnell and other researchers who worked during the 1960s on the biochemistry of memory deserve a place in the pantheon for the founders of the modern search for the engram because their work was a bridge that connected the older nonphysiological tradition of the comparative psychology of invertebrate animal learning with modern developments in biochemistry and molecular biology. The legacy of McConnell and the other planarian researchers is the establishment, with very well-controlled experiments, of classical conditioning in invertebrates.

The Search for the Engram Before McConnell and the 1960s

To appreciate McConnell's successful struggle to establish learning in invertebrates and his contribution to neuroscience in the area of the biochemistry of memory, it is necessary to consider the state of knowledge about invertebrate learning and the physiological basis of memory around 1950. Lashley is cited frequently by contemporary researchers in neuroscienee who work on the cellular basis of learning and memory, but McConnell is largely forgotten. Lashley's failure by 1950 to localize the engram or memory trace at a place in the nervous system (see Donegan & Thompson, 1991; Finger, 1994) led to speculation that the engram was biochemical (McConnell, 1967a).

Urged on by a suggestion in Hilgard's (1948) classic text on learning theory, researchers in animal learning and comparative psychology were searching for an invertebrate preparation as a launchpad for the physiological study of learning because invertebrate nervous systems have fewer neurons than vertebrate systems. Unfortunately, there was a major conceptual stumbling block. The conventional wisdom, especially among zoologists and others who were not experts in animal behavior, was that invertebrates were little robots without an internal state for memory, in which behavior was guided by instincts. Even Maier and Schneirla (1935), in the leading textbook of comparative psychology, described invertebrate learning as ephemeral when they wrote,

Experience may temporarily alter the form of behavior by inducing local tissue change.., but such changes are wiped out by subsequent events, and have no permanent altering effect.

p. 84.

However, by citing a study on planarian learning from the 1920s, Maier and Schneirla left open the possibility that their generalization about invertebrate learning might not apply to planaria.

Richard Thompson, the senior author with James McConnell (Thompson & McConnell, 1955) on a nowclassic study on learning in planaria, studied with Lashley. Thorne (1995) designated the little-known and publicityshy Thompson as Lashley's heir. Because Lashley failed to find the engram with rats, Thompson and McConnell were motivated to investigate learning and memory with an invertebrate preparation, so as to succeed where Lashley had failed.

The Search for the Engram Before McConnell and the 1960s

During the 1960s, James McConnell was one of psychology's most visible and colorful public personalities, a celebrity-scientist who made entertaining appearances on television. As the head of the Planarian Research Group at the University of Michigan, McConnell's research program was a precursor of the kind of interdisciplinary approach to brain and behavior that is now called neuroscience. He wrote,

This prospect of working with a covey of bright scientists in disciplines other than Psychology pleases us immensely

McConnell, 1967a, p. 2.

His approach encompassed measurement at multiple levels of investigation. As he put it:

✓ If learning is (at one level of discourse) a matter of some kind of structural or functional change at the synapse, shouldn't we also wonder what chemical change takes place at the synapse when 'learning' occurs?.. antireductionism is misplaced;
✓ Each scientific discipline surely has something unique and important to contribute to the solution of the problem of memory formation;
✓ all the disciplines are equally important.

McConnell, 1967a, p. 2.

Usually a field honors its pioneers. Yet McConnell and many of the other scientists who pioneered the biochemistry of learning and memory during its modern, formative period in the 1960s have become nonpersons - eclipsed, put down, or written out of the contemporary story of the search for the engram. Alport (1986), a science writer, called this omission citation amnesia. Although McConnell wrote an annual review of invertebrate learning in 1966, the most recent annual review of invertebrate learning by Krasne and Glanzman (1995), which included references to the late 19th and early 20th centuries, contained no primary citations of McConnell and the work from the 1960s on planarian learning. McConnell's (1966) article contained 109 references, none of which survived for citation by Krasne and Glanz. Why has this work virtually disappeared from contemporary citation?

Memory Transfer: McConnell's Blind Alley

One reason for the missing citations to McConnell is that his memory transfer paradigm was a failure. In these cannibalism studies, which McConnell (1962) saw as a technique for transferring a memory molecule of RNA from trained to untrained organisms, a naive planarian showed savings in the acquisition of a conditioned response (CR) when fed the body parts of a planarian that had learned a classical conditioning task. McConnell's research program with planaria collapsed when other scientists failed to replicate the phenomenon of memory transfer. The failure of memory transfer has probably overshadowed McConnell's success with invertebrate learning. Because others (Collins & Pinch, 1993; Donegan & Thompson, 1991; Rose, 1992; Travis, 1980, 1981) have told the colorful story of the failure of the memory transfer research, this article concentrates on McConnell's successful struggle to establish the study of invertebrate learning as a respectable endeavor.

McConnell's Struggle With Critics Over Invertebrate Learning

Today invertebrate learning is well established (Abramson, 1994) as a robust, long-lasting, ecologically valid phenomenon (Krasne & Glanzman, 1995), presupposed by the scientists who work on the molecular dissection of memory (DeZazzo & Tully, 1995). In McConnell's day, the critics simply did not believe that invertebrates could learn.

The original Thompson and McConnell (1955) study, a demonstration of Pavlovian conditioning in planaria, is a classic article on invertebrate learning (Abramson, 1994). It launched McConnell's career. Planaria live in water and normally glide along the bottom of pools on slime trails they lay down. The classical conditioning trials were administered while the planarian was gliding in water from one end of a foot-long trough to the other. Rather than recording the CR automatically, Thompson and McConnell followed a more ecological tradition of comparative psychology and used naturalistic observation to score contraction. For the experimental group, which consisted of the pairing of a light from above with shock through the water, the percentage of trials with a contraction CR increased modestly from about 2% during the first 50 trials to 10% during the last 50 trials. Another response, turns which were more common, increased from a high baseline of about 25% to a conditioned rate of 35%. For the three control groups (light alone, shock alone, and a naive group with neither a CS nor an unconditioned stimulus [US]), the rate of contractions did not change.

Thompson and McConnell were graduate students of M. E. Bitterman at the University of Texas. Bitterman, a distinguished comparative psychologist, had studied with Schneirla, so these students inherited a great tradition of comparative psychology. McConnell's first struggle over planarian learning was with his mentor, Bitterman, and it occurred before the article was even submitted for publication. Bitterman (1975), who wanted a control group with unpaired presentations of light and shock, left us with a very critical commentary on the scientific skills of his neophytes. Here is Bitterman's retrospection:

Is it really so difficult to understand that CS-alone and US-alone groups do not, even together, control for the effects of stimulation per se on an experimental group exposed to both s t imul i?.. Not only are the controls employed in the search for invertebrate learning generally inadequate, but the techniques are crude and subjective.

Bitterman, 1975 p. 140.

Bitterman's critique raises an impor tant question about the conflict between novice innovators and established scientists. Creative - especially young - scientists may want to publish innovative findings as rapidly as possible to establish their reputations, sometimes even before the technique is perfected and all of the relevant variables are understood. More established scientists, whose reputations are secure, want to protect the literature from unreplicable phenomena and poorly controlled experiments. Here, the conflict was resolved in favor of publication by Harry Harlow, editor of the Journal of Comparative and Physiological Psychology. History shows that Harlow made the right decision.

In 1955, Harlow wanted to publish articles using nontraditional (i.e., nonrat) species because of a famous article by Beach (1950) called The Snark Was a Boojum. Beach's point was that the journal was becoming a journal of rat psychology, so he called for a greater variety of species. A game that a sophisticated reviewer for a journal can always play with a novice investigator who is moving a line of work in a new direction is to demand additional control groups that are different from those used in the article under review. By relaxing the editorial criterion for originality, some flaky articles will be published, but then other scientists can sort out the issues by running replications that include the inevitably necessary controls.

The controversy about planarian learning highlights the constructive role of the scientific critic. McConnell, an innovator, raced from one exciting phenomenon to the next without comprehensive experimental analysis or adequate controls. McConnell's controls were often developed as a response to his critics. McConnell's students and other scientists were left the task of cleaning up after McConnell by adding the control groups that he omitted. After his arrival at the University of Michigan, Donald Jensen became McConnell's nemesis over invertebrate learning - Jensen's position was

that no invertebrate, no matter how complex is capable of showing 'true learning'

McConnell & Shelby, 1970, p. 75.

Jensen (1965) attributed the results ofplanar ian learning to sensitization and called for better control groups. McConnell met Jensen's objections by upgrading the quality of the control groups. As he put it,

There isprobably no other group of scientists as enamored of the use of 'control groups' as are psychologists; indeed the entire history of scientific psychology may be viewed as a continuing search for better controls

McConnell, 1967a, pp. 25 - 26.

The planarian learning controversy brought passion to the discussion of control groups for classical conditioning because a Nobel prize was thought to await the discovery of a biochemical engram.

Controlling for Experimenter Bias

Experimenter bias was one of the simplest criticisms. Cordaro and Ison (1963) manipulated the expectancy of students in a course in introductory psychology who were running an experiment in classical conditioning on planaria. Sure enough, the students who were led to expect contractions reported more responses than students who were led to believe that conditioning would not occur. McConnell (1967a) quickly instituted a blind control procedure in which the experimenter did not know the group to which the planarian was assigned, so experimenter bias failed to explain the data.

Controlling for Pseudoconditioning and Sensitization

Baxter and Kimmel (1963) repeated the original Thompson and McConnell (1955) study with the addition of an unpaired control group. The unpaired group equated experimental and control groups for the amount of exposure to the CS and US. For the classical-paired group, the number of trials with a CR increased to 50%, whereas the unpaired group showed a steady decrease in responding during conditioning. To distinguish between pseudoconditioning - the enhanced responding to a CS that is not dependent on a forward, temporal CS-US relationship - Jacobson, Horowitz, and Fried (1966) added a backward conditioning control group. There was no learning with the backward conditioning group. Ultimately, by giving the animals only a few trials per day, lengthening the intertrial interval, and not running the animals each day, McConnell (1964) was able to obtain a CR on 90% of the trials. Thus, the controversy about planarian learning produced a steady improvement in the quality of the data as the researchers identified optimal parameters.

Today, a common control procedure in invertebrate learning is a discrimination in which a CS+ is paired with the US, and a CS- is presented alone. This procedure controls for sensitization, an increase in responding to a CS that does not depend on the forward pairing of the CS with the US. This control was introduced to invertebrate learning with planaria by Block and McConnell (1967) to address concerns raised by Jensen (1965) about sensitization as an alternative explanation to associative learning. Block and McConnell established exactly the kind of discrimination called for by Jensen by implementing an elegant A - B - A - B reversal design withinsubjects. One CS was vibration produced by a speaker mounted below the trough, and the other CS was the traditional illumination from lights mounted above the trough. Paired presentations in Phase A increased responding to CS+ but not CS- , whereas extinction in Phase B reduced responding. As Block and McConnell concluded in their study,

The findings of the present study.., should go a long way towards answering in the affirmative the question, 'Can planarians be conditioned reliably?'

p. 1466. Amen.

Even after he was well established as a comparative psychologist at the University of Michigan, McConnell was sometimes unable to obtain respect for his work on planarian learning from scientists in other disciplines. Libbie Hyman, a zoologist, was the world's leading expert on invertebrates. After establishing that invertebrates could learn, McConnell moved on to the problem of memory. McConnell, Jacobson, and Kimble (1959) demonstrated that planaria could retain an association established by Pavlovian conditioning for four weeks. During the early 1960s, when McConnell visited Hyman at her office in the American Museum of Natural History in New York City, Hyman dismissed McConnell's article on retention with the following words,

I'm very sorry, but I just can't believe that. No that just can't be. I could believe that a planarian might remember something for five minutes or so. But weeks or months? No that just can't be.

McConnell, 1970 p. 1.

Clearly, the expertise of comparative psychologists in behavior analysis had not yet earned the full respect of scientists in other disciplines during the 1960s. Some simply did not believe the data.