The Discovery of Insulin

by Michael Bliss

  35 Macleod 1922/78, appendix, Macleod to Banting, March 11, 1921. Compare Banting 1940, p. 24: “But he said that no one could look after the dogs and no one was interested, so I had better wait….”

  36 My interviews, the Hipwell memoir, and Harris’s account all support the view that the engagement was broken off before Banting insisted on coming to Toronto, and therefore not because of his determination to work on the internal secretion. But that determination, when it became clear, probably did not improve relations, for it was a sign of Fred’s continued unwillingness to settle down.

  37 Stevenson 1946, p. 73; see also Dale 1946, pp. 37–38.

  38 Banting 1940, p. 18.

  39 Ibid. But the date is open to question. In his 1922 statement Banting does not mention a trip to Toronto on April 26, and implies (using in his rough draft the same image of turning the key and taking a lone suitcase) that he made his trip on May 14. It is natural, however, that Banting would have taken a few days in Toronto as a break from teaching and to firm up the details of the work with Macleod. The commencement date of April 14, 1921, given in his Nobel Prize lecture (Banting 1925) is clearly wrong.

  40 BP, 1, Starr to Banting, May 3, 1921.

  41 Macleod 1922/78. As Macleod remembered it, they talked about using extracts. The Academy Notebook, however, suggests that they probably also discussed using transplants of degenerated pancreas.

  42 Macleod 1922/78, note 3; Collip 1916, 1920, 1921. There is some unclarity about which meeting Collip attended. It might have been the June meeting when Macleod gave Banting the suggestions recorded in Banting’s notebook on June 9 and 14.

  43 The wording of Macleod’s statement (1922/78, note 5) may be further evidence that an assistant had not been talked about earlier: “I gave Dr. Banting all the assistance I could in planning the details of the experiments, in searching the literature and, finding that he was entirely unfamiliar with the chemical methods necessary for the proposed investigation, I offered him the assistance of one of my research fellows, C.H. Best, who had been trained in this work.”

  44 See E.C. Noble’s unpublished account, dated October 1971, in the Noble Papers for the statement that Macleod introduced Banting to Best and Noble together. Best 1922 simply refers to meeting Banting early in May (Best also says that in the lecture Macleod mentioned that Banting’s idea might lead to the development of “an efficient pancreatic extract”). Later (Best 1972), Best thought that he and Banting met in the autumn of 1920 and at that time discussed the work Banting wanted to do in Macleod’s laboratory. There is no evidence of this meeting in any of the 1922 accounts, including Best’s, although it would be natural for either Banting or Best to mention it. No other account, including those of Best, refers to an earlier meeting.

  45 Banting 1940, p. 25.

  46 Noble account, 1971. Macleod 1922/78 apparently contradicts this, referring to Best only. But Best 1922, supported by both Noble and Banting, makes clear that he and Noble were to divide the time.

  47 In most of his later accounts of the discovery (the exception is one interview, Stalvey 1971), Best did not remember the coin-tossing incident.

  Banting 1940, p. 25, mentions the desire to avoid a break in the summer holidays.

  The sources contain a bewildering set of contradictory statements about how many weeks the assistants were to work. Some accounts say four weeks each, others say three. Banting gave both versions. In one of Best’s accounts he refers to four weeks and two weeks. And so on.

  The confusion stems from the fact that only six weeks remained between mid-May and the usual beginning of holidays, but Banting proposed to work for eight weeks or more. The final arrangement is not clear. There were many possible ad hoc arrangements that the trio could have worked out. Perhaps Best and Noble were each to give Banting three weeks’ time in a four-week period. Perhaps Banting thought he would not need help during or after a certain period. The fact that Best knew he would be going off for the last two weeks in June for militia training may have further complicated the arrangements.

  Best raised another issue in later accounts of the beginning of the work by stating that he worked without pay and that he considered himself to have volunteered to work with Banting. See, for example, the statements in Best, Selected Papers, p. 5; also Best 1972; also Wrenshall, et al. 1962, p. 61.

  While it is true that Best later volunteered to do Noble’s stint, and probably did it without thought of payment (though Macleod afterwards arranged that he and Banting were paid for their time), all of the 1922 sources clearly point to the conclusion that Best was assigned to assist Banting as part of fellowship duties for which he was being paid. He and Noble received an extra $85 cash for their work in June and knew they would be fellows in 1921–22 at a salary of $800 each (MP, Macleod to Falconer, May 31, 1921; Macleod to F.A. Mouré, June 3, 1921). There is no evidence that Banting, although heavily in debt and short of ready cash, asked for, expected, or was offered payment, according to the original arrangement. He was the volunteer outsider with the idea. Best was his paid assistant.

  48 Banting 1922, rough draft.

  49 This uncertainty, plus the impossibility of dating Banting’s discussions with Macleod, makes it impossible to state precisely when the Toronto work that led to the discovery of insulin formally began. Banting left London on Saturday, May 14. He began work in Toronto on either Monday, May 16, or Tuesday, May 17. May 17 was the date of the first experiment.

  50 Banting 1940, p. 25.

  Chapter Three: The Summer of 1921

  1 Unless otherwise noted, all details concerning the animal experiments are taken from Banting and Best’s original notebooks. These consist of Banting’s original notebook at the Academy of Medicine in Toronto, and the several joint notebooks used by Banting and Best which are in the Banting Papers at the University of Toronto. There are also some loose notes accompanying the original charts of the experiments in the Banting Papers. Day-by-day summaries of the experiments which I compiled as background to these chapters have also been deposited in the Banting Papers.

  2 See Hédon 1909 for a description of the operation. In several of his accounts Best mentions that the first job was to look up this and other literature.

  3 Academy Notebook; Best 1942, 1974; FP, 1921 file, notes of a conversation with C.H. Best, Jan. 20, 1956.

  4 It is not clear whether Macleod was responsible for this suggestion. Best 1922 states that Macleod had suggested a D:N ratio higher than they ever obtained, but whether it was the very high 3.65 figure (which was common to animals made glycosuric by a very different chemical procedure, but well above the 3.00 rough average for depancreatized dogs) is unclear. See also Best 1942: “Banting had understood from Professor Macleod” that 3.65 was necessary for complete diabetes. There could easily have been a misunderstanding.

  5 The sources do not disclose whether they planned to use autogenous or heterogenous grafts, or both.

  Dr. F.C. Macintosh, professor emeritus of physiology at McGill University, who read an early draft of this manuscript, suggests a plausible, slightly more speculative, variation of my interpretation of the research plan:

  As I see it there weren’t to be two approaches, but a single one in which each stage, if successful, led logically to the next. This was to be the basic sequence: First, to demonstrate, by planting a working Hédon remnant into a diabetic animal, that one dog’s internal secretion could prevent another dog from getting diabetes. That accomplished, to show that a ground-up pancreatic remnant could do, for a brief period, what the intact remnant had done over a longer period. Finally, to replace the emulsion (which I guess would have been pretty toxic if given by vein) by one or other kind of extract: the difference between an aqueous emulsion and an aqueous extract being essentially that the latter had been cleaned up by filtering or centrifuging it to remove the larger particles. This stagewise approach would reflect Macleod’s caution. It would ensure that the most critical observations would be made on dogs that had been kept
in good shape for some time after subtotal pancreatectomy, rather than on dogs that might be deteriorating from a combination of causes – surgery & infection & diabetes. And it might delay the final push to get the hormone cleaned up and bottled until the surgery had become routine, Macleod had returned from Scotland, and Collip was available to help with the biochemistry. As a sop to Banting, there would be quite a lot of varied surgery. And probably Macleod agreed that if the transplants failed repeatedly, the experimenters could try emulsions or extracts.

  6 Margaret Mahon Best ms, “The Discovery of Insulin,” privately held; FP, ms biography of Best.

  7 That second 386 was a puzzling dog. It had not healed quickly after its first operation on May 31, occasionally showing a high blood sugar. After the second operation, apparently completing the pancreatectomy, its blood sugar stayed in the mildly diabetic range of .20 to .30, its D:N, after the first surge, became insignificant, and on the day it died, June 27, its blood sugar was a normal 116. The autopsy showed open wounds, but no infection and no trace of pancreas. It is most likely that both the operation and the autopsy missed portions of pantcreas – notoriously easy to miss in both procedures – which had sustained partial pancreatic function in the dog.

  8 Banting 1940, pp. 27–28.

  9 Best 1942, p. 389, refers to “a great deal of discussion of the results of the glucose and nitrogen estimations.” Also FP, Best to Feasby, May 6, 1957, in which Best writes (p. 11), “I persuaded Fred with difficulty that Macleod was wrong about D/N ratios. This threatened to wreck our partnership because Fred was a stubborn man. He eventually thanked me repeatedly and warmly.”

  10 Banting 1940, p. 26. Some of the phrasing used in the contemporary accounts and some oral accounts leads to a possible interpretation that Noble simply did not appear at the beginning of July, that Best filled in for him, and only then decided to stay on.

  11 Banting 1940, p. 45; Cody Papers, G.W. Ross ms, “History of the Discovery of Insulin,” undated but probably 1941.

  12 Catgut is not mentioned specifically in the notebooks, but is referred to in Best 1942. The notebooks do show that silk was used for the second series of ligations. In the Cameron Lecture (Banting 1929), Banting gave: a more detailed and slightly divergent account: “We chloroformed a couple of the dogs which had their pancreatic ducts ligated…. Careful examination showed that the ligature was still present in a bulbous sac in the course of the duct. It was therefore necessary to operate on all the duct-tied dogs a second time and to exert particular care as to the tension put on the ligature. If the ligature was applied too tightly gangrene developed immediately underlying the ligature and a serous exudate laid down on the surface over the ligature resulted in the recanalisation of the duct. If applied too loosely the duct was not blocked. We, therefore, in some cases applied two or three ligatures at different tensions.”

  13 Best’s letters to Margaret Mahon were not available for this study. But the references to the work in the lab in these letters have been excerpted and are quoted in Margaret Mahon Best, “The Discovery of Insulin,” and also in two summaries compiled by C.H. Best and W.R. Feasby in the Feasby Papers, 1921 file. There are minor dating errors in the summaries.

  14 BP, 1, Best to Macleod, Aug. 9, 1921; for the preparation of the extract see the notebooks, also Banting and Best 1922A, 1923, Banting 1923A.

  15 Banting 1910, pp. 25–26; Banting and Best 1922A; Banting 1929.

  16 BP, 1, Best to Macleod, Aug. 9, 1921.

  17 Ibid.; also chart in BP. The notebooks are riddled with errors concerning this dog. Its first-stage operation was on July 22. There is no record oí a second stage, and on July 30 Best noted that dog 106 was dead. And yet 406, severely diabetic, receives extract on Aug. 1. It appears that the second stage was done on 406 but not recorded, and that the dog recorded as dead on July 30 was actually 107, which had been done on July 26 and never again appears in the records.

  18 Best 1922.’

  19 BP, 1, Banting to Macleod, Aug. 9, 1921.

  20 Notebooks. Banting and Best 1922A incorrectly lists the doses as live cc.

  21 See note 13 above.

  22 BP, 1, Banting to Macleod, Aug. 9, 1921; BP, 22, index card list of “Ideas,” dated Aug. 8. Banting also listed “Alcohol extraction” on the card, but did not put it in the list for Macleod.

  23 BP, 1, Banting to Macleod, Aug. 9, 1921.

  24 Details from Ross manuscript, Hipwell 1970, interview with Mrs. Fannie Lawrence.

  25 Hipwell 1970.

  26 Note 13 above; also notebooks. The rabbits are not referred to in the notebooks.

  27 Banting 1940 and various Best accounts, particularly those in FP, National Film Board files; also Hipwell 1970.

  28 Banting 1940, pp. 39–44; also Banting 1929. An account of an encounter with an anti-vivisectionist regarding this dog in Banting 1940 does not seem a plausible occurrence at that time; it probably happened a year or two later.

  29 Banting and Best 1922A.

  30 Another factor contributing to the error stemmed from their having made three preparations of whole gland extract. The neutral extract, given on August 17, had the marked effect. Acid extract, given early the next morning, had a slight though still noticeable effect; alkali extract had none. By making the three separate preparations they had precluded giving serial injections, although this had been their common method. For discussion of the hypothesis, see pp. 203–8.

  31 Notebooks; Banting and Best 1922A. In the first secretin experiment they also stimulated the vagus nerve to try to obtain more pancreatic juice after secretin stimulation.

  32 Banting 1929, 1940.

  33 Banting 1940, p. 42.

  34 Banting 1940, p. 18. It is not clear where Banting got a car. His 1940 account seems to show that he did not have a car when he came to Toronto and that the story of his selling his old Ford that summer to raise money is inaccurate. He actually sold his beat-up old Ford in the autumn of 1920. Of course he might have bought another car by August 1921. More likely, he had borrowed one, perhaps the university’s “Pancreas car,” referred to later.

  35 Interview with Dr. Ian Anderson.

  36 BP, 1, Macleod to Banting, Aug. 23, 1921.

  37 The notebooks show some evidence of this. In their first paper Banting and Best promised a more detailed description of the histological sections in a subsequent communication: “Suffice it here to note that the pancreatic tissue removed after seven to ten weeks’ degeneration shows an abundance of healthy islets, and a complete replacement of the acini with fibrous tissue” (Banting and Best 1922A). No records of these sections can be found and nothing more is said in any subsequent communication. None of the tissue had been subjected to more than seven weeks’ degeneration.

  38 Macleod’s letter must have seemed so out-of-date that Banting, a year later, apparently told people he had never received the letter. See Macleod 1922/78, Note 4.

  39 Compare Banting 1940, p. 37: “I have always advocated that no person should be allowed to operate on a dog, who has not received training as a surgeon.”

  40 Perhaps significantly, they never again identify their extract as “Isletin.” The term is not used anywhere in their notes, or in any other source after August 9.

  41 The mess with cat extract had caused them to resolve to centrifuge all future extracts (this is the only change introduced from their original methods of preparation, although some of the doses of extract were being made slightly acid), but they made their first injection of this new extract before it had finally settled. The dog had a marked reaction. It is not clear whether centrifuging was regularly used.

  42 Banting 1940; Macleod’s return is noted on Banting’s 1922 desk calendar, BP, 26.

  43 Stevenson 1946, p. 89; the notebooks indicate bowel problems with dog 92; Noble papers, Robert Noble interview with E.C. Noble, April 12, 1977; interviews.

  44 Personal communication.

  45 Falconer Papers, 71, Henderson to Falconer, Sept. 21, 1921; Banting 1940, p
. 29.

  46 Macleod 1922/78; Best 1922.

  47 Banting 1922. Also Macleod 1922/78: “I found that Banting and Best were dissatisfied with the facilities at their disposal, and early in October they formally demanded of me that I improve them.”

  48 In Banting’s 1922 account he implies that Macleod relented and made specific commitments before the meeting ended. In his 1940 account the change of heart is not evident until the next morning. Macleod 1922/78 is not helpful on this point, but mentions Banting’s threats to go to the Mayo Clinic or the Rockefeller Institute.

  49 FP, National Film Board file, Best dictation, transcript March 20, 1956.

  Chapter Four: “A Mysterious Something”

  1 Banting 1940, p. 31; Banting 1922.

  2 Falconer Papers, 71, Henderson to Falconer, Sept. 21, 1921; Macleod 1922/78, appendix, Macleod to Falconer, Sept. 30, 1921. In his memoirs and afterwards Banting warmly thanked Henderson for providing this position, thereby, Banting claimed, making it possible for him to stay in Toronto. There is no direct discussion of this job arrangement in the sources, but the documents give the impression that at some point Henderson consulted Macleod about the idea of appointing Banting in Pharmacology. Macleod would have supported this enthusiastically as an excellent way to keep Banting in Toronto without having to put up a fight with the administration, which he might not win, for a special position and extra money for his own department. There is no evidence that Macleod ever turned Banting down for a job in Physiology.

  3 Banting 1940, pp. 35, 49–50.

  4 Macleod 1922/78. Banting 1940, p. 46: “I asked Macleod if Collip could join us and work on the biochemistry aspect but Macleod said there was ‘plenty of time for that later on if necessary’.” Banting 1922: “I was very anxious that the work advance more rapidly. I asked Professor Macleod three or four times if Dr. Collip could do portions of the work, but he advised against it.”

  5 BP, 26, desk calendar 1922. In his 1922 account Banting says Connaught gave them three calves for exhausted gland experiments. There are no records of these.