The Discovery of Insulin

by Michael Bliss

  6 Notebooks. In Banting and Best 1922A the ligation period is incorrectly stated as six weeks.

  7 Notebook notations on October 5 and 9, read in conjunction with the note on the sources of the extracts in Banting and Best 1922A, lead to no other interpretation than this. That this complete confusion had occurred cannot be realized from the publication alone; in it the relative strengths of the different extracts are stated, but not the relative degrees of degeneration of poor Towser’s pancreas. It is just possible that an error in the notebooks is corrected in the publication.

  8 BP, 22.

  9 BP, 22. This is the only index card in Best’s handwriting in the group. There is always the possibility, of course, that Banting did the reading and dictated the note to Best.

  10 Paulesco 1923A.

  11 This is the second part of the error in Banting and Best 1922A, and it does not follow from either Best’s index card or the Paulesco article. There is a bare possibility that a missing second index card, plausible given the ending of the back of the first card, might have contained further Best notes on Paulesco leading to the error.

  12 BP, 22, index cards. The article was McGuigan 1918.

  13 BP, 1, Banting to Starr, Oct. 19, 1921.

  14 BP, 26, desk calendar 1922.

  15 Several people to whom Banting told his version of the discovery recalled how emphatically he mentioned Macleod having made them repeat their work. The most explicit reference is in Macleod’s 1926 account, which seems to refer to exactly this stage in the history of the experiments: “In view of the large amount of work which had previously been done in this field, it was considered advisable to make certain of the anti-diabetic effects of the extracts, as judged by the behaviour of blood sugar and urinary sugar, before proceeding to investigate their influence on other symptoms of diabetes, such as glycogen formation, ketosis, and changes in respiratory metabolism” (p. 65). The various unpublished accounts in the E.C. Noble Papers also stress Macleod having made them repeat their summer’s experiments.

  16 Scrapbooks of Mrs. C.H. Best, Scrapbook No. 1.

  17 Banting and Best 1922A.

  18 BP, 39, typed diary entry, Nov. 14, 1921.

  19 Banting 1922.

  20 Macleod 1922/78.

  21 Macleod 1922/78 is the only source giving this origin of the idea of the longevity experiment. But no other account contradicts this or gives any other version of its origin. I assume that “H.B.” Taylor in Macleod’s account is a typographical error for N.B. Taylor, a member of the Department of Physiology at that time.

  22 Banting’s accounts of his reasoning correlate unusually smoothly. See Banting 1922, 1925, 1929, and Banting and Best 1922B, 1922C. Undated inde’x card notes on Laguesse and Carlson & Drennan are in Banting Papers, 22. In most accounts Banting mentions that only later did they find an article by Ibrahim showing that trypsin/ogen/ is not present in the foetus during the first third of pregnancy. It would not be surprising if his memory was incorrect on this last point.

  23 Banting and Best 1922B, 1922C.

  24 Banting and Best 1922B.

  25 See Banting and Best 1922B for the first precise statement of quantities: “we placed 50 grams of tissue in 250 cc. of saline, macerated and filtered; 15 cc of this solution we’re then diluted to 250 cc. with saline. A 15 cc dose of this solution reduced the percentage of blood sugar in a 10 kilogram dog from .40% to .15% in three hours.” The context indicates this was done on November 20–22. The notebooks do not record these quantities, nor any injection showing these results.

  26 Banting and Best 1922B.

  27 Banting and Best 1922C.

  28 The chart in Banting and Best 1922C contains all the data compiled on dog 27, with only one error. The experiment was being very poorly conducted, and it would be difficult to show that the pancreatectomy, despite autopsy findings of no pancreas, had actually made the dog diabetic. On Dec. 2, for example, a blood sugar reading 69 hours after the last injection of extract was .15.

  29 Macleod 1922/78.

  30 My summary chart of the results is now in the Banting Papers. I am grateful to Professor O. Sirek for suggesting this approach to the early results.

  31 Macleod 1922/78.

  32 MP, folder marked 342, Joslin to Macleod, Nov. 19, 1921; reply Nov. 21.

  33 BP, 2, Clowes to Banting, Sept. 25, 1934.

  34 Notebooks and Banting and Best 1922C. The latter seems incorrect in dating the dog’s second reaction as December 4, rather than December 3.

  35 The use of this name for dog 33 seems to have arisen only some years later. There are no contemporary references.

  36 Best 1922.

  37 For preparation methods see Banting and Best 1922B, 1922C. For the use of warm air see Macleod 1922/78. The description of the injection to dog 33 in Banting and Best 1922C, p. 6, contains quite incorrect figures. The first injection of 11 cc. on dog 33 did not cause a fall from .30 to .15 in one hour; the actual drop was from .35 to .30. The impressive drop in the first 45 minutes, from .26 to. 14, had actually taken plate the day before on dog 23.

  38 Banting 1940, p. 46; also Banting 1922, in which he mentions having persuaded Velyien Henderson to intercede with Macleod, and Macleod having told Henderson, in effect, “that the scientific world would think he was silly if he gave up his work on Anoxaemia for the investigation of the extract, since its success had not yet been proved.” Banting places this discussion in January, but it must have been earlier.

  39 Collip Papers, Medicine, transcript of K.K. Shouldice statement, October 1, 1958. The statement contains some factual errors and I am not sure of its reliability.

  40 Falconer Papers, 71, V.J. Harding to Falconer, Sept. 27, 1921; on Collip generally, see Barr and Rossiter 1973; for his love of research, see Collip Papers, Medicine, Ms of Addison Lecture, July 19 18.

  41 FP, 1922 file, Best, “The Discovery of Insulin,” Osier Oration-July 12, 1957. See also FP, “Hot” file, Best to Feasby, May 6, 1957; and FP, Best to Sir Henry Dale, Feb. 22, 1954. In his 1910 account, p. 48, Banting alludes to Best being suspicious of Collip at an early stage in the collaboration.

  42 I have located five documents by Collip describing his contributions to the insulin work. There are two published articles, Collip 1923K, 1924; an undated, signed statement, “The Contribution Made by J.B. Collip to the Development of Insulin While be Was in Toronto 1921–22” in his papers in the medical library at the University of Western Ontario; and two undated second pages of letters written from the University of Alberta, both apparently written in 1923. One of these is in a collection of Collip papers privately held; the other was found in what were then unsorted files of Collip papers in a storeroom in the Biochemistry Department at the University of Western Ontario. Collip also considered that Macleod 1926 contained a true account of the discovery.

  In his signed statement Collip dates the rabbit experiments as beginning on December 12.

  43 See Macleod 1922/78 for his claim to have made the suggestion, which Collip then acted upon. This is not contradicted in the literature except by Best 1922, “there is considerable doubt as to whom belongs the credit of the idea.” He adds, however, “Dr. Collip was first to put the idea into effect.”

  44 Banting and Best 1922 C gives an incorrect number to the dog, an incorrect volume of extract, and mis-states the extent of the decline in blood pressure. It is also almost certainly incorrect in stating that whole beef extract was used, for the notebooks seem to show that the first preparation of this extract was not made until Dec. 12. The type of anesthetic used in the experiment is not stated.

  45 Collip’s accounts all omit the Dec. 9 experiment by Banting and Best on the anesthetized dog and convey the impression that the work on the liver began with Collip’s experiment on the normal dog. Possible explanations of this omission are that Banting and Best’s experiment was so inconclusive as to be virtually meaningless, and/or that the train of thought about the liver, set out in the text, did not occur until Collip repeated the e
xperiment. The notebooks show, however, that Banting and Best did remove specimens of dog 23’s liver for glycogen estimates, and in their paper 1922B, though not 1922C, they present the liver hypothesis as following from this experiment. There is no record, on the other hand, of the glycogen estimates actually having been done after this experiment.

  My ordering of the evidence seems consistent with all of the accounts, and particularly Macleod’s 1922/78 summary: “It was agreed that the time had now arrived when it would be advisable to ascertain the glycogen content of the liver and other viscera in depancreated dogs treated with the extract. Dr. Collip undertook to do this and to observe the excretion of acetone bodies in these animals prior to death.” Collip later found out that potent batches of extract did lower the blood sugar of anesthetized dogs.

  46 Banting 1910, 1922; Best 1922. The dog they refer to must be the Airedale described in Collip 1923K. There may have been another dog experimented on at the same time.

  47 Macleod 1922/78. Best 1922: “I had been desirous of performing this experiment at an earlier date, but had been unable to obtain the apparatus.”

  48 BP, 26, desk calendar.

  49 The notebook lists Dec. 16; in 1922C and on the chart the date is Dec. 15.

  50 A month or so later it became a matter of dispute, as we will see, as to who had first discovered the insolubility of insulin or an insulin-compound in 95 per cent alcohol. In their second published paper, written in February, Banting and Best mention that this extract of Dec. 15 or 16 was also washed in 95 per cent alcohol. This washing is not recorded in the notebooks. Nor is there any record in the notebooks or the published papers of a procedure, described by Banting in both his 1922 and 1940 accounts, in which – in Collip’s presence in the 1940 but not the 1922 account – 95 per cent alcohol was added to an 80 or 70 per cent alcoholic solution of the active principle and the solution became opalescent, a very fine precipitate developing.

  Nor is there any record in the notebooks or the publications of an experiment first described in Best and Scott 1923D, p. 712, and attributed to Banting and Best, in which, after evaporation and washing with toluene, “they treated an aliquot portion of the dried residue with 95 per cent alcohol. The mixture was filtered and the filtrate evaporated. The residue was dissolved in saline solution. Administration of this solution produced no effect upon the blood or urinary sugar of a depancreatized dog. A saline solution of the material which did not dissolve in 95 per cent alcohol, however, definitely lowered the blood sugar and diminished the sugar excretion of the same animal.”

  The notebooks do show that on December 21 Banting and Best tried an extract “of the crystal first deposited by the alcohol ext. drying in warm air current.” This memory, of a precipitate formed at a point where the solution was nearly pure alcohol, may have formed the basis of Banting and Best’s claim to have discovered the insolubility. Except that the precipitate proved inactive in the test animals.

  The other possibility is that the experiment described by Best and Scott was done early in January. But the complete absence of notes in Banting’s otherwise apparently quite complete sequence, and of any references in the 1922 publications, is puzzling.

  51 Gilchrist, Banting, and Best 1923.

  52 BP, 22. In his 1922 account Banting also mentions this test, dating it December 21 and indicating that the mode of testing was for glycosuria.

  53 Collip 1923K; also Collip Papers, Medicine, “Contribution of J.B. Collip to the Discovery of Insulin…”

  54 See Collip 1923K, p. 8, for details of this experiment. The ketone readings are in Banting, Best, Collip, and Macleod, 1922B.

  55 Collip 1923K. See also Macleod 1926, pp. 101–102: “the percentage of glycogen found in the liver after the animal had been given very large quantities of sugar along with insulin was so large that it was difficult to determine with accuracy; it was apparently over 20 percent.” This wording, combined with a more general neglect of this result in Macleod’s accounts, suggests that he may have doubted Collip’s result with this first glycogen estimation.

  56 Collip 1924.

  57 In his 1922 account Banting states that Collip told Macleod the results of this experiment, implying that he and Best only learned later, presumably on the train. This became another item in Banting’s indictment of Collip and Macleod. It might well have happened, perfectly innocently, if Collip ran across Macleod before or during Christmas week. In any case Collip, who considered Macleod the director of the work, would naturally tend to report to him first.

  Chapter Five: Triumph

  1 Banting, Best, and Macleod 1922.

  2 Banting 1940, p. 48.

  3 Macleod 1922/78. Joslin 1956 makes the same point in a more kindly way: “The possibility of mistakes in the work was fully exploited by those who discussed the paper in a skeptical but on the whole a sympathetic way.” Another eye-witness wrote later in 1923: “Whether it was the excessive modesty of the speaker, or whether due to a somewhat apathetic attitude on the part of the audience, Dr. Banting’s address made little impression on the members present.” Funk and Harrow 1923. In 1972 R. Carrasco-Formiguera recalled the meeting this way: “I remember my amazement and disgust at some whom I had previously held as giants, but who now tried to underrate such a remarkable achievement. I also recall my pleasure at others, including those who themselves had been close to success, who acknowledged the achievement of the young Canadians, as well as the originality and importance of their findings.”

  4 Joslin 1956.

  5 See E.L. Scott to Macleod, Feb. 7, 1922, printed in A.H. Scott 1972, pp. 136–38, in which, referring to their discussion at New Haven, Scott mentions the “marked reactions” Macleod had described the Toronto preparations as causing.

  6 FP, Best dictation, transcribed March 20, 1956.

  7 See the argument in Pratt 1954.

  8 See Roberts 1922, Pratt 1954, and pp. 203–8 of this account.

  9 FP, 1924 file, Feasby interview with Joslin, Nov. 1957.

  10 MP, Allen to Macleod, Feb. 8, 1922; Scott to Macleod, Feb. 7, 1922, in A.H. Scott 1972, pp. 136–8.

  11 Macleod 1922/78; Clowes 1948.

  12 Banting 1922. By 1940 Banting had convinced himself that he could have handled the questions, and describes the meeting this way: “When I sat down there was considerable discussion and many questions asked. I noted each and since the discussion was from the audience and not from the platform, I had forgotten all about myself and was prepared to talk freely in reply to questions and discussions. To my surprise, however, I was not called upon by the chairman as was the invariable rule. Macleod himself responded to all questions and expostulated theories and referred to ‘our work’ and ‘I believe’ and ‘I think’.”

  13 Sec ch. 4, note 57.

  14 Banting 1940, p. 49.

  15 MP, Macleod to A.R. Cushing, Jan. 7, 1922; Macleod 1922/78.

  16 In Banting and Best 1922C they state it got 6 cc. of whole gland extract daily from December 8. The notebooks record, however, two injections of foetal calf extract, totalling 16cc, for December 8, 15 cc. of foetal calf extract on December 9, 10 cc. whole gland extract on December 15 and 16. There are no records of other injections and no charts. On December 11 the notebooks record that the dog’s picture was taken. The pictures printed in Banting and Best 1922C are captioned as being nine weeks after total pancreatectomy.

  17 Banting and Best 1922C.

  18 Best 1922 states that all work after Christmas was carried out under Macleod’s direction. Both Banting and Macleod in their 1922 accounts refer to the more formal division made in late January.

  19 Best 1922: “In the winter of 1922 I spent most of my time in superintending the collection and initial concentration of material which was then handed over to Dr. Collip for completion.”

  20 The only explicit statement that this was Banting’s expectation is in Best 1956.

  21 See Collip 1923K, 1924; Banting, Best, Collip, Macleod, and Noble 1922A, 1922C.

 
22 Gaebler 1965.

  23 Noble Papers, Noble accounts, March 12, 1977, October 1971; also Robert Noble interview with E.C. Noble, 1977, and author interview with Robert Noble 1980.

  24 Mann and Magath 1921. Collip’s accounts of these observations imply that he discovered the reaction and its antidote independently. If so, he was surprisingly out of touch with the most current literature.

  In a Feb. 22, 1954 letter to Sir Henry Dale, (FP) intended for posthumous publication, Best states that “Banting and I had noted hypoglycemia and had recorded this and the beneficial effects of sugar.” There is no note or record of this.

  25 Macleod 1922/78.

  26 Tory Papers, File 504–9, Collip to Tory, Jan. 8, 1922.

  27 Banting 1922.

  28 Interview with R.B. Kerr, 1980, who is writing a biography of Graham.

  29 Banting 1940, p. 55.

  30 Macleod 1922/78.

  31 In his Feb. 22, 1954 letter to Sir Henry Dale (FP), Best describes the background as follows: “Banting began to get very restless about Collip’s activities and his relationship to us…and spoke to me one day about preparing some potent insulin to give to the first human case of diabetes which was to be treated on the wards of the Toronto General Hospital. Banting’s statement to me was: ‘The insulin which Collip is making may be somewhat freer from impurities than that which you have made and which we have given to depancreatized dogs. We know, however, that the material which you have made from whole beef pancreas, is really potent and that it gives no obvious reaction in the dogs, i.e. no local reaction.’ He said: ‘I think it would be much more appropriate, Charley, in view of our work together, if this first case should receive insulin made by your hands and tested by us on dogs and on ourselves.’ I had no hesitation in agreeing….”

  32 Interview,’ July 10, 1980.

  33 Allan 1972.

  34 Campbell 19 16, 1962; Banting, Best, Collip, Campbell, and Fletcher 1922.

  35 This account of the preparation of the extract relies on Banting and Best’s published description in Banting, Best, Collip, and Macleod 1922A. There are variations in the description Best gave in his Feb. 22, 1954 letter to Dale. In later accounts of the first clinical test (e.g., Best 1956), Best stressed that he would have much preferred to use foetal pancreas, that he had said this to Banting at the time, but was persuaded that they had to use a commercially available source. Had extracts of foetal pancreas been used, Best seemed to be saying, they would have been more successful. Such an extract was not used. Whether or not it would have been more successful is not known. Best sometimes cited clinical tests done many years later which showed the potency of foetal pancreas (Salter, Sirek, Abbott, and Leibel 1961), but a close examination of the method of extraction used in that study shows it was not identical to that published or noted by Banting and Best in 1921–22.