Repost: Musty Must-Read: "A study of trial and error reactions in mammals"

Recent discussion of the way papers should be presented and comments on the way papers were written in the good old days when Uncle Sol was a wee scientist motivated me to repost something I put up on the old blog July 11, 2007.


First, I'll tip the hat to Shelley at Retrospectacle for starting a "tour of the vaults" with the classic LSD in elephants study. Today, I'm reaching way back for "A study of trial and error reactions in mammals" by G. V. Hamilton, Journal of Animal Behavior, 1911 Jan-Feb 1(1):33-66. This study is worth reading because it provides an often hilarious insight into the conduct of science at the turn of the past century but also because this study is a root (perhaps the taproot) of a relatively current subfield on spatial working memory and spatial search.


The studies describe a spatial search task in which the subject enters at one end of the apparatus (ENT D on the figure) and then can choose one of 4 doors to exit the apparatus. The monograph is communicated from "Montecito, California", undoubtedly from apparatus set up at the author's ranch, see below. Each trial presents only a single unlocked door and sequential trials never repeat the same unlocked door. The goal of the study was to examine search patterns in a fascinating range of subjects, the first instructive and hiliarious aspect of the study. A subset:

"Man 1 Age, 34 years Native (Spanish-Indian) Californian Ranch laborer in this experimenter's employ. A man of limited education, but of average intelligence for his class. He went through his trial in the stolid, unemotional manner that characterized his work in the fields. The "boss" wanted him to walk into and out of an enclosure 100 times, and he did so without asking questions or shirking his task"

"Boy 6F1 Age, 13 years. Father Italian, mother Swedish. Country schoolboy. He was less alert, mentally, than were his brothers, who are described below"

"Defective Man A. Age, 45 years. ....Ranch laborer in the experimenter's employ... nervous suspicious, "muddled" person, with a grievance against society in general, and a surprising fund of self-acquired misinterpretations relating to social environment. He expressed a belief that my experiment was dangerous meddling with the human mind... constant dread of apparatus...labored under a suspicion that it was not the simple structure that it pretended to be"

"Monkey 3. Age, 5 years (est). Macacus rhesus (sic). About 3 years in captivity. A truculent, untameable animal, but a fairly good subject"
" [Ed: under Dogs section] ...descended from a common sire, the subject of my previously published 'An experimental study of an unusual type of reaction in a dog'"

"Cat 1. Age, 1 year. Manx. This animal was reared in the Harvard Psychological laboratory and enjoyed the further distinction of having been one of Doctor Berry's subjects... "

"Horse 2. Age, 8 years... in view of the very poor showing made by this animal, the stableman's belief in the "smartness" is of some interest."

The absolute best quote of the whole paper:

"it is much to be regretted that the list given below includes no human subjects between the ages of 26 months and eight years; the writer's wholly undeserved local reputation as a vivisectionist seemed to create a stubborn unwillingness on the part of parents to supply young children for experimental work" (emphasis added)

Hamilton1911-Fig1
I won't detail the findings exhaustively. Suffice it to say Hamilton essayed a comparative cognition study on search behavior, trying to stratify his subjects by some analog of "smartness". Due to the heterogeneity of his subject population, it ended up being....rather subjective. Nevertheless, this is the instructive rather than hilarious part. It was a decent attempt, undertaken prior to the establishment of many of the components of behavioral research we take for granted at present time. Many of the scientific issues are still being explicated and the technical observations presage modern methods. One of his human subjects sounds suspiciously schizophrenic, for example. Humans will work for cash, monkeys for appetitive reinforcers. Monkeys are "highly distractable" but otherwise good subjects. The human developmental comparisons attempt to account for both "nature" and "nurture" components (albeit in a manner that wouldn't pass muster today). It is not simple to get parents to allow you to use their children in research!
As I mentioned above, this paper is at root of considerably more modern investigations. A series of mid 80's papers on memory effects of lead exposure and scopolamine challenge by Levin and Bowman actually include "Hamilton Search Task" in the title. The Wisconsin National Primate Center also includes this task in a standard battery. More recently, the "self-ordered spatial search" task (aka, "spatial working memory task") of the CAmbridge Neuropsychological Test Battery (CANTAB) has obvious roots in the Hamilton task. This task has been used in many studies, human as well as animal; as a brief subset- studies of frontal lobe damage in human, related marmoset studies here, and rhesus here and here).
Science is an incremental business, built slowly on the brickwork of often distant generations of scientists.

3 responses so far

  • Frasque says:

    That really was funny, thanks for sharing.

  • Now *this* is why I enjoy the old literature - not just for the colorful explanations (and the rampant IRB and IACUC violations) but also for the fact that this ranch-based experimentation formed the basis of modern work and is still referred to as the Hamilton Task.
    I have to admit that I had to look up "truculent." I don't see how the monkey could be truculent and a "fairly good subject" at the same time.
    1911, wow. Thanks for reposting this - I had missed it at the old blog.

  • becca says:

    All right. You've convinced me. I want to see more supplementary data files that have descriptions of subjects like that.

Leave a Reply