This view is still held by some, but others, in view of the recent discovery of Deichmüller and Tollens,³⁶ that diabetic urine when distilled yields decidedly more acetone than alcohol, have suggested that the substance is derived from aceto-acetic acid.

³⁶ Loc. cit.

The first test suggested for acetone was Gerhardt's chloride-of-iron test. A solution of chloride of iron added to urine containing acetone strikes a burgundy-red color. But this reaction occurs with so many substances that it cannot be considered entirely reliable. Ralfe's modification of Lieben's iodoform test³⁷ is made as follows: About a fluidrachm (3.7 c.c.) of liquor potassæ, containing 20 grains (1.2 grams) of iodide of potassium, is placed in a test-tube and boiled; a drachm (3.7 c.c.) of the suspected urine is then carefully floated upon the surface. When the urine comes in contact with the hot alkaline solution a ring of phosphates is formed, and after a few minutes, if acetone or its allies are present, the ring will become yellow and studded with yellow dots of iodoform, which, in turn, will sink through the ring of phosphates and deposit itself at the bottom of the test-tube. A number of other substances produce the iodoform reaction, but only one of these, lactic acid, is likely to be met in urine.

³⁷ Clinical Chemistry, Philadelphia, 1884, p. 100.

The perspiration, saliva, exudations, and effusions in diabetic cases have all been found, at times, to contain sugar.

DURATION.—Diabetes is a disease of which the duration is measured by months and years, and although cases are reported in which death supervened in from six days to six weeks after the recognition of the disease, it is evident that such periods do not necessarily measure its actual duration. The disease may have existed some time before coming under observation. On the other hand, a case is reported by Lebert which lasted eighteen years; another, under the successive observation of Prout and Bence Jones, sixteen years; and a third, under Bence Jones and Dickinson, fifteen years. The younger the patient the shorter usually is the course run and the earlier the fatal termination. Yet I have known a girl of twelve recover completely. After middle age the disease is usually so easily controlled by suitable dietetic measures, if the patient is willing to submit to them, that its duration is only limited by that of an ordinary life, while carelessness in this respect is apt to be followed by early grave consequences.

COMPLICATIONS.—The almost sole complication of diabetes mellitus is the tubercular phthisis which so often terminates it. Indeed, it is doubtful whether this complication should not be regarded as a consequence, as should also the boils, gangrenous processes, and ophthalmic conditions which have been mentioned under Symptomatology. Jaundice has occurred three times in my experience up to the present time. Senator says that when not an accidental complication due to a catarrh of the duodenum it may result from compression of the biliary capillaries by the overloaded blood-vessels and enlarged gland-cells of the liver. In one of my cases, in which jaundice appeared to be the initial symptom, but which disappeared some months before death, the autopsy revealed atrophy of the liver. It is well known that pancreatic disease, especially cancer, is apt to be accompanied by jaundice, and as pancreatic disease is often at the bottom of diabetes, it will similarly account for the jaundice, while the presence of jaundice may also suggest a pancreatic diabetes.

DIAGNOSIS, INCLUDING THE TESTS FOR SUGAR IN THE URINE.—The diagnosis of diabetes mellitus, the disease being once suspected, is easy. The passage of large amounts of pale urine of high specific gravity, the presence of thirst, dryness of the mouth, fauces, and skin, and progressive emaciation even while the appetite is good, can scarcely be misinterpreted. In the urine from such a case the application of any of the tests for sugar will produce prompt response. The urine is not always so much increased as to attract attention, while its color is also sometimes but slightly changed; but the symptoms of thirst and dryness or clamminess of the mouth are seldom wanting. On the other hand, the discovery of a glycosuria without these symptoms is, as a rule, accidental. It is a question how far such degrees of glycosuria as do not produce the usual symptoms of diabetes in an appreciable degree are signs of positive disease. At the same time, its detection is important, in that there is always danger of the simple glycosuria becoming a diabetes—a danger which its recognition and suitable treatment may avert. Accordingly, the urine of all persons having unusual appetites without evident cause, and of those who are fond of eating and drinking, should be tested for sugar. This should also be done for those who have passed through severe mental or physical strain, have suffered shock or concussion of the nervous system, blows upon the abdomen, etc.

Testing for Sugar.—Under the head of Diagnosis I prefer to include the testing for sugar, which requires some detailed consideration. Unless it be that the indigo test recently revived by George Oliver of London prove more delicate, that form of cupric test known as Fehling's solution is, with suitable precautions, all things considered, the most satisfactory for general use.

Fehling's volumetric solution, suitable for both qualitative and quantitative purposes, is made as follows: Dissolve 34.639 grams of pure crystallized cupric sulphate in about 200 cubic centimeters of distilled water; 173 grams of chemically pure crystallized neutral sodio-potassic tartrate and 80 grams of potassium hydrate in 500 or 600 c.c. of distilled water. To the latter add the copper solution slowly, and dilute the clear mixed fluid to 1 liter. One cubic centimeter of this solution will be decolorized by 0.005 grm. of sugar, or 200 grains will be decolorized by 1 grain of sugar. Or the copper may be dissolved in 1 liter of water, and the tartrate and potassium hydrate in another, and a cubic centimeter of each mixed at the moment they are to be used.

For qualitative testing, put a cubic centimeter of Fehling's solution into a test-tube (or if the copper and the alkaline sodio-potassium tartrate solutions are kept separate, a cubic centimeter of each), and dilute with distilled water to 5 c.c. Boil, and if, after the lapse of a couple of minutes, the solution remain unchanged, it is fit for testing. If it becomes turbid or a red sediment falls, it is spoiled, and a new solution should be obtained.³⁸ A cubic centimeter of the suspected urine is then measured out and added drop by drop to the solution kept hot. If there is much sugar, the first drop will throw down a yellow precipitate of suboxide of copper, which becomes rapidly red. If no reaction takes place after adding the entire cubic centimeter of urine, the addition should be continued until 4 c.c. are added, when, if, after the mixture has cooled, there be no response, it may be concluded that the urine is free from sugar. By operating with a cubic centimeter of the test-fluid and the same quantity of urine or multiples thereof, we may roughly estimate the proportion of sugar. Thus, if the cubic centimeter of undiluted urine just decolorizes the cubic centimeter of Fehling's solution, sugar is present in the proportion of one-half of 1 per cent.; or if a half cubic centimeter of the urine removes all the color, the quantity is 1 per cent. If the urine is highly charged with sugar, it may be diluted, and the degree of dilution being remembered, a rough quantitative estimation may be similarly made.

³⁸ Should this not be possible, a little more soda may be added and the fluid filtered, when it is again ready for use.

If the urine contains very minute quantities of sugar, the reaction is less satisfactory. The copper is reduced, but the suboxide is so small in quantity that it is obscured by the excess of copper solution, and a mixture results which is greenish or greenish-yellow or yellow or milky, and on standing a small yellow sediment falls to the bottom. Now, it dare not be said that it is sugar which produces such reaction. It may be sugar, but it may also be uric acid. Uric acid is really more frequently a source of error than is commonly supposed. I have myself seen the reaction due to it so vivid that I did not suspect it could be due to any reducing agent excepting sugar; but, noting the next day a copious sediment of uric acid which had fallen during the night, a testing of the supernatant fluid then revealed no reaction whatever. Such a urine, after being treated by the lead process to get rid of the uric acid, fails also to respond. But this process is very tedious,³⁹ and cannot be conveniently carried out by the busy practitioner. The same thing is, however, accomplished by treating the urine with hydrochloric acid, which in twenty-four hours precipitates all of the uric acid. Simple precipitation by lead acetate solution and filtration does not answer, because all of the uric acid is not thus removed. Other substances, as hippuric acid, urates, hypoxanthin, etc., are said to act similarly, but they produce no practical interference with the test. On the other hand, a small amount of sugar may be present and yet fail to show the reaction, because the cuprous oxide is held in solution by certain substances. Such are ammonia and nitrogenous matters, including albumen, creatinin, pepsin, peptones, urinary coloring matters, etc. The latter probably produce their effect through the ammonia which is given off while heating them in the presence of an alkali. Hence all albumen should be precipitated and filtered out of urines suspected to contain sugar, and the heat applied should not be too great. Finally, excess of glucose will also hold in solution cuprous oxide, so that the suspected urine should not be added in too large a quantity at a time, but rather drop by drop.

³⁹ The details of this process will be found in the writer's work on the Practical Examination of Urine, 5th ed., 1883, p. 63.

But qualitative testing is not sufficient during the treatment of a case of diabetes. The percentage of sugar and the quantity discharged in twenty-four hours should be determined occasionally. The process is done as follows: Place 10 cubic centimeters of Fehling's solution in a porcelain capsule, and dilute it with 40 c.c. of distilled water. Fill a Mohr's burette with the urine, which, if it contain more than 1 per cent. of sugar, should be diluted with nine times its bulk of distilled water. Slowly heat the contents of the capsule to boiling, and then allow a little of the diluted urine to run in from the burette; continue the cautious addition of urine and the gentle heating until the blue color is completely removed from the Fehling's solution. To determine the exact moment at which this takes place requires a little experience, but its recognition is facilitated by carefully tilting the capsule after each addition and stirring, so that its clear white surface may be seen through the edge of the fluid and contrasted with the latter. The number of cubic centimeters of urine used should now be read off from the burette, the number of c.c. of undiluted urine calculated therefrom, and each c.c. multiplied by .005 grm. The result indicates the quantity of sugar in grams in the urine employed, whence the percentage of sugar is determined, and also the twenty-four hours' quantity, the amount of urine passed in that period being known.

The Fermentation Test.—A very simple and easy method of determining the proportion of sugar is by Roberts's fermentation method, which, although not so precise as the volumetric process, is still sufficiently so for clinical purposes. A small piece of German yeast or a teaspoonful of liquid yeast is added to about four ounces (120 c.c.) of the urine, which is kept lightly stopped, at a temperature of 20° to 30° C. (68° to 80° F.), for about twelve hours; at the end of this time the sugar will have been converted into alcohol and carbonic acid. The latter will have passed off, and the urine lost in weight because of the destruction of sugar; while the difference between the specific gravity before and after the fermentation indicates the number of grains of sugar per fluidounce. Thus, suppose the specific gravity before fermentation to have been 1040, and afterward 1025; there will have been 15 grains of sugar to the fluidounce, whence, again, the twenty-four hours' quantity can be calculated. If the metric system is used, each degree of specific gravity lost will correspond to .2196 grams of sugar in every 100 c.c. of urine.

The specific gravity of the fermented urine should be compared with that of the urine soon after it is passed, because saccharine urine under suitable circumstances undergoes fermentation without the addition of yeast; and, the specific gravity being thus lowered spontaneously, the reduction in the urine fermented by yeast would appear less than it actually is. At the same time, care should be taken that the urine is of the same temperature when the specific gravity is taken before and after fermentation.

The Picric Acid and Potash Test.—Although attention was called in 1865 by C. D. Braun,⁴⁰ a German chemist, to a reaction between grape-sugar and picric acid, as the result of which the latter is converted into picramic acid, very little attention seems to have been paid to this announcement. Quite ignorant of it, George Johnson rediscovered this reaction in 1882, and published it in 1883.⁴¹ It is applicable to both qualitative and quantitative purposes. In order to make use of it, a standard comparison-solution is made as follows: Take 1 fluidrachm of a solution of grape-sugar, 1 grain to the fluidounce; mix it in a long test-tube with half a drachm of liquor potassæ (U. S. P. or B. P.) and ten minims of a saturated solution of picric acid; dilute the mixture to 4 fluidrachms with distilled water, to facilitate which a tube used for the purpose may be marked at 4 fluidrachms. Raise the mixture to the boiling-point, and continue the boiling for sixty seconds, to ensure complete reaction between the sugar and picric acid. During the boiling the pale-yellow color of the liquid is changed to a vivid claret-red. Cool the liquid by cautiously immersing the tube in cold water, and if it is not then at the level of the 4-drachm mark, raise it to this by adding distilled water. The standard color thus obtained is that which results from the decomposition of picric acid by a grain of sugar to the ounce, four times diluted, or by a solution of sugar containing one-quarter of a grain per ounce. But the picramic solution rapidly becomes pale on exposure, so it becomes necessary to make a more permanent solution to use as a standard. This may be accomplished by combining liquor ferri perchloridi drachm j, liquor ammonii acetatis drachms iv, acidum aceticum (glacial) drachms iv, and water enough to make ounces iiss. The color of this is identical with that of the picric acid reduced by a one-grain solution diluted four times, and, according to Johnson, it will retain its color unchanged for at least six months. At the same time, whenever a new solution is made it should be compared with that of the one-quarter grain per ounce solution of sugar, boiled with picric acid and potash.

⁴⁰ "Ueber die Umwandlung der Pikrinsaüre in Pikramminsaüre, und Ueber die Nachweisung der Traubenzucker," Zeitschrift für Chemie, 1865.

⁴¹ British Medical Journal, March, 1883.

For qualitative testing Johnson directs: To a drachm of urine in a test-tube add a few drops, enough to give a distinct yellow color, of a saturated solution of picric acid. Add about 10 drops of liquor potassæ and boil. If sugar is present, the mixture becomes promptly red in hue.

It is plain, then, that if a given quantity of the dark saccharine fluid produced by boiling—say, enough to cover ten divisions of the graduated tube, as shown in the figure—has to have added to it an equal bulk of distilled water in order to produce the color of the standard solution, the tested fluid will be of the strength of two grains to the ounce; if three times, three grains; and so on; while fractional additions, as indicated by the graduated markings, would show fractional additions to the proportion of sugar.⁴²

⁴² A more exact comparison of the saccharine liquid with the standard is made by pouring into a flat-bottomed colorless tube six inches long and an inch in diameter as much of the standard solution as will form a column about an inch in height, and an exactly equal column of the saccharine fluid in a precisely similar tube. The operator then looks down through the two tubes at once, one being held in each hand, upon the surface of a white porcelain slab or piece of white paper. In this way slight differences of tint are easily recognized; and if the liquid to be analyzed is found darker than the standard, it is returned to the graduated tube and diluted until the two liquids are found to be identical in color, when the final reading is made.

The presence of albumen, even in considerable amount, has but little effect upon the test, nor does the coloring matter of normal urine, according to Johnson; but he says there is a coloring matter associated with ser-albumen in albuminous urine, and with egg-albumen as well, which has a reducing action on picric acid. This is partly separated by filtering off the precipitated albumen, and entirely removed by repeated filtration through animal charcoal. So, too, the albumen removed by coagulation and filtration, if thoroughly washed, does not give any red reaction if boiled with picric acid and potash diluted in the same proportion as when testing for sugar. Neither do any other unoxidized sulphur compounds found in urine decompose the picric acid and render the test fallacious.

Johnson and his son, G. Stillingfleet Johnson, claim that the picric-acid test is as accurate as any other, and that it is even more accurate than either Fehling's or Pavy's process, because the picric acid is not acted upon by uric acid or urates, which do reduce the oxide of copper. The method of analysis by the picro-saccharimeter, they claim, is at least as speedy and as easy as any other. The materials and apparatus required are easily prepared, inexpensive, and not, like Fehling's copper solution, liable to undergo rapid changes.

But while Johnson claims that neither coloring matters of normal urine nor uric acid reduce the picric acid, he admits that he has tested with picric acid and potash a large number of specimens of normal urine with the almost uniform result of a depth of color indicating the proportion of .6 of a grain of sugar to the fluidounce, the indication varying between the limits of .5 to .7 grain. The ammonio-cupric method used at the same time gave results of from .7 to .9 grain to the fluidounce, or an excess of .1 to .3 grain. Now, if my own views, the grounds for which are announced elsewhere,⁴³ are correct, strictly normal urine contains no sugar, and any reducing action upon oxide of copper is due to uric acid, either picric acid is reduced to a degree by uric acid or by some other constituent of normal urine. This, in the light of Oliver's⁴⁴ recent investigations, may be kreatinin. For he has shown that kreatinin strikes in a few seconds a red color with the cold alkaline picric solution, which is quickened by heat. From this it would seem that the exact value of the picric-acid test has as yet to be determined.

⁴³ Tyson, Practical Examination of Urine, 4th ed., Philadelphia, 1884.

⁴⁴ On Bedside Urine-Testing, including Qualitative Albumen and Sugar, by Geo. Oliver, M.D., London, Member of the Royal College of Physicians of Lond., etc., 2d ed., London, 1884.

The Indigo-Carmine Test.—The fact that indigotine, the coloring matter of commercial indigo, is converted into indigo when heated with an alkali in the presence of glucose and certain carbohydrates, has recently been applied by George Oliver of London in the construction of a test-paper. Carmine of indigo is the sulph-indigotate of sodium, an intensely blue salt, soluble in 120 parts of water. Sulph-indigotic acid is made by heating indigo with sulphuric acid, and when combined with a base, sodium, produces indigo-carmine. When sodium carbonate is mixed with a solution of indigo-carmine, the latter is precipitated in a minute state of division, but is redissolved on heating, when there results a greenish-blue solution. A freshly-made mixture of the indigo solution and sodium carbonate furnishes a fluid not unlike Fehling's solution, which gives the reaction to be described with glucose. Unfortunately, such a mixture will not keep, and the reagent would be useless but for the happy idea of Oliver of making the test-paper. In doing this bibulous paper is immersed in a solution of indigo-carmine with carbonate of sodium.⁴⁵ The paper is then cut into strips an inch long and one-quarter of an inch wide.

⁴⁵ No more precise directions than this are given by Oliver, either in his papers in the Lancet for 1883 or in his little book just published, On Bedside Urine-Testing. The sugar test-papers, as well as the entire series of albumen test-papers, suggested by Oliver, are now made by Parke, Davis & Co. of New York, and by Wilson & Son, Harrogate, London.

Mode of Testing.—One of the test-papers and a sodium carbonate paper⁴⁶ are dropped into a half-inch test-tube, and water added until the upper end is just covered; a column of fluid one inch in height and half an inch in diameter will thus be produced, so that the solution of carmine obtained on boiling will always acquire the same concentration. Heat is now applied, the tube being gently shaken, and boiling kept up for a second or two. A beautiful blue solution will result. The test-paper may now be removed or allowed to remain.

⁴⁶ Test-papers of the same size, charged with a saturated solution of sodium carbonate.

Not more than one drop of the suspected urine is let fall into the tube from a pipette held in an upright position. Drops of equal size are thus secured. The contents of the tube are again freely boiled for a few seconds, after which the tube should be raised an inch or more from the flame and held without shaking, while the solution is kept quite hot, but not boiling, for exactly one minute. If glucose be present in abnormal amount, the soft rich blue will be seen first of all to darken into violet; then, according to the quantity of sugar, there will appear in succession, purple, red, reddish-yellow, and finally straw-yellow. When the last-named color has been developed the slightest shaking of the tube will cause red streaks to fall from the surface and mingle with the pale yellowness of the solution, while further agitation will cause the return of purple and violet and the restoration of the original blue.

The time required for the commencement of the reaction after the boiling of the test liquid is in inverse proportion to the amount of glucose present. When the latter is large, over 20 grains to the ounce, it will be but a few seconds; but when small, 2 or 3 grains, from thirty to sixty seconds may elapse. If the urine do not contain more than the normal amount of sugar⁴⁷—i.e. under half a grain to the ounce—the color of the solution at the end of the heating for one minute will be unchanged. The test is available by artificial light as well as by daylight.

⁴⁷ It will be noted from this that Oliver accepts the view that there is a small amount of sugar in normal urine.

Precautions.—1. Care should be taken during the testing not to shake the tube or to permit free ebullition. 2. While keeping the contents of the tube hot, the latter should not be held up between the eye and the sky, for then the early color-changes will probably escape observation. The tube should be kept below the eye-level and its contents viewed by the reflected light of some bright object, such as a sheet of white paper propped up an inch or two beyond the tube as a background. 3. Oliver is not aware that the presence of earthy carbonates will prevent the carmine reaction, but as a precautionary measure he suggests the use of a soda-paper whenever the water is exceptionally hard. 4. The acids of the urine rob the carmine-paper of much alkali, so that the addition of more than a certain number of drops of urine—varying of course with the degree of acidity—will at first retard and then prevent the reaction. The addition of the soda-paper will prevent any such interference, although Oliver says that by invariably submitting only one drop of saccharine urine to the test-paper, and keeping up the heating for not less than two minutes, he has never failed to obtain the characteristic reaction without using a soda-paper. It is well to remember, however, that an excessively acid urine may thus interfere, and that the soda-paper will prevent it. 5. The blue color of the carmine is discharged by caustic alkali—liquor potassæ or sodæ. The only chance of being misled by this reaction lies in using an imperfectly cleansed test-tube which may have contained Fehling's solution or the alkaline picric solution. The caustic alkali converts the blue carmine into a green solution, which, on heating, disappears; nor does it return by again shaking the contents of the tube.

Critical comparison of this test with Fehling's solution and picric acid by Oliver has shown that of sixty-four substances experimented upon, normal and abnormal constituents of urine or medicines which after ingestion are eliminated in the urine, Fehling's was reduced by fifteen, picric acid by eleven, and indigo-carmine by eight. The only substances producing the characteristic play of colors with indigo-carmine test-papers reacted with both picric acid and Fehling's solution. They were unoxidized phosphorus, ammonium sulphide, milk-sugar, dextrin, inosit, gallic acid, tannic acid, and iron sulphate. Both the carmine and picric acid were reduced by inosit, which merely turned Fehling's solution green. On the other hand, uric acid and urates, which reduce Fehling's solution, do not react with the carmine test, while kreatinin, which reacts with picric acid also, does not respond to the carmine. Albumen, if abundant, interferes with Fehling, but not with the indigo-carmine.

Detection of Inosit.—It has been said that inosit sometimes accompanies, and even substitutes, grape-sugar in the course of diabetes. It has been mentioned that it does not reduce Fehling's solution, but turns it olive-green. It reduces the carmine and alkaline picric acid solution, and is therefore not recognizable by these. The methods recommended for its recognition in the books are troublesome, and as its presence in the absence of sugar indicates a favorable change, it is not likely that a more precise recognition than is furnished by the olive-green reaction will be needed for clinical purposes.

PROGNOSIS.—The prognosis in diabetes depends upon the organ whose involvement is responsible for the symptoms, upon the stage at which the condition comes under observation, and upon the age of the patient. It has appeared to me that the cases of diabetes depending upon pancreatic disease are the most intractable, that their progress is scarcely checked by treatment, and that they are comparatively rapidly fatal in their termination. In the others, where the symptom is one of a central nervous lesion, it has always seemed to me to be of secondary importance that the glycosuria is itself less marked, that it is unattended by the other distinctive symptoms of diabetes, and that its issue is that of the nervous malady.

Again, it is well known that the later in life diabetes occurs the more amenable it is to treatment, and that if a proper diabetic diet be adhered to by the patient his life need scarcely be shortened. On the other hand, diabetes mellitus is a disease in which the expectant plan is dangerous. If it does not improve it usually gets worse; and many a patient has fallen a victim to his own indifference and indisposition to adhere to a regimen under which he could have lived his natural term of life. This is especially the case when the disease appears after middle life.

If, on the other hand, the condition becomes thoroughly established before twenty-five years of age, it is less amenable to treatment; but even in such cases a promptly vigorous treatment is sometimes followed by recovery. I have already mentioned the case of a child twelve years old in which complete recovery took place.

If tubercular phthisis supervenes, recovery is not to be expected, while intercurrent disease, as pneumonia, which is rather prone to occur, is very much more serious and apt to terminate fatally.

TREATMENT.—The treatment of the aggregate of symptoms known as diabetes mellitus is conveniently divided into the dietetic, the medicinal, and the hygienic, of which the first is by far the most important. The efficiency of this treatment depends upon the successful elimination from the diet of all articles containing grape-sugar, cane-sugar, beetroot-sugar, and starch, it being universally recognized that in the early stages of the disease these foods are the sole source of the glucose in the urine. The normal assimilative action of the liver, by which the carbohydrates are first stored up as glycogen, and then gradually given out as glucose or maltose to be oxidized, being deranged, such foods not only become useless as aliments, but if continued seem to aggravate the glycosuria, and the excretion of sugar steadily increases. There is, therefore, a double reason for excluding them from the food. This is easiest accomplished by an exclusive milk diet. The exclusive milk treatment of diabetes was suggested by A. Scott Donkin in 1868. That he is correct in his assertion that in the early stages of diabetes lactin or sugar of milk is quite assimilable, and does not in the slightest degree contribute to the production of glycosuria, I cannot doubt; that it is in this respect even superior to casein, as claimed by Donkin, I am not prepared to state from actual knowledge; but that casein itself resists the sugar-forming progress immeasurably greater than any other albuminous substance, so that in all but the most sure and advanced or complicated cases its arrest is complete, I am also satisfied. Certain it is that in a large number of diabetics the use of a pure skim-milk regimen results in a total disappearance of the sugar from the urine. That in a certain proportion of these cases a gradual substitution of the articles of a mixed diet may be resumed without a return of the symptoms is also true. In other more confirmed cases the use of skim-milk results in a decided reduction in the amount of sugar, with an abatement of other symptoms, which continues as long as the diet is rigidly observed. In still other cases, while the skim-milk treatment makes a decided impression upon the quantity of sugar, it still remains present in considerable amount, while the disease progresses gradually to an unfavorable issue. These three classes of cases represent, ordinarily, different stages of the disease, so that it may be said that as a rule cases recognized sufficiently early may be successfully treated with skim-milk, although it may occasionally happen that cases pursue a downward course from the very beginning despite all treatment. Yet I have never seen a case which, when taken in hand when a few grains of sugar only to the ounce were present, failed to yield to this treatment.

As to the method of administration, my practice with adults is to give eight ounces (an ordinary tumblerful) every two hours, beginning at seven or eight o'clock in the morning, and continuing to the same hour in the evening. Sometimes it is well to begin with half as much at first, but rapidly to increase to the required amount. This method ensures the ingestion of three to four quarts daily—a quantity generally sufficient to maintain the body-weight of an adult person of average size and taking moderate exercise, although a slight reduction may take place at first. But if the individual is very active or of large size, it will not be found sufficient. In such event the quantity must be increased as demanded by a feeling of unsatisfied hunger. I have known fourteen pints to be taken in twenty-four hours. But when the quantity becomes thus large, the inconvenience in ingesting it is very great, and it is much more convenient to coagulate the casein of a part of the milk and use the curd thus obtained, while the second part is drunk. Curd may be seasoned with salt to make it more palatable, and should be thoroughly masticated before it is swallowed.

The milk should not be taken too cold, especially if the amount ingested is large, else it is likely to reduce the temperature of the stomach below the point necessary for gastric digestion. The temperature should not be less than 60° F., nor much over 100°. Something depends upon the idiosyncrasies of the patient, which must be the guide as to temperatures intermediate between those named.

The chief advantage of the skim-milk over the unskimmed is simply that it is more easy of digestion. Many persons who cannot take unskimmed milk for any length of time without its deranging the digestion, or, as is commonly said, making them bilious, can take with impunity milk from which the cream is removed. Although Salomon⁴⁸ claims to have shown that glycogen is produced in the liver of rabbits fed upon pure olive oil, it is at least probable that fat is among the last of the substances undergoing this conversion, and in ordinary cases of diabetes it is rather its indigestible nature which renders it prudent to remove from milk the greater proportion of fat by skimming it off.

⁴⁸ Virchow's Archiv, Bd. 61, Heft 3, 1874, 18.

Still more easily assimilable is the peptonized milk, in which the casein is at least partially digested, and it should be employed where there is any difficulty in the way of using the ordinary milks. Either skimmed or unskimmed milk may be used for peptonizing, the latter peptonized being quite as easy of digestion as the former unpeptonized. I have found the extractum pancreatis of Fairchild Brothers & Foster most successful in the peptonizing of milk, and according to the following directions: Into a clean quart bottle put 5 grains of extractum pancreatis, 15 of bicarbonate of sodium, and a gill of cool water; shake, and add a pint of fresh cool milk. Place the bottle in a pitcher of hot water or set the bottle aside in a warm place, usually for three-quarters of an hour. When the milk has acquired a slightly bitter taste, it has been completely peptonized—that is, the casein has been completely converted into peptone. After the process is complete the milk must be immediately put on ice.

It is not always necessary to completely peptonize the milk, and if the bitter taste is unpleasant the process may be stopped short of this by putting the milk on ice, the degree of digestion depending upon the length of time the milk is kept warm.

While I am confident that the promptest and most effectual method of eliminating sugar from the urine is by a milk diet, it occasionally happens that a patient cannot or will not submit to so strict a regimen. In other instances, again, it is not necessary to resort to it, because a less restricted diet answers every purpose.

A suitable diabetic diet would also be obtained by eliminating from the bill of fare all saccharine and amylaceous and other sugar-producing substances. Such a diet is, strictly speaking, impossible. For, apart from the fact just mentioned that even fats, as well as albuminous substances to a degree, are capable of producing glycogen, the monotony of a pure meat diet soon becomes unbearable, to say nothing of other derangements it may produce. Fortunately, it is not necessary that such an exclusive diet should be maintained, for certain saccharine foods seem capable of resisting the conversion into sugar more than others. Sugar of milk, or lactin, has already been mentioned as one of these, and to it may be added the sugar of some fruits, and probably also inosit or muscle-sugar, mannite or sugar of manna, and inulin, a starchy principle abundant in Iceland moss. It is found also that there are many vegetable substances containing small quantities of sugar and sugar-producing principles which may be used with impunity in at least the milder forms of diabetes. This being the case, a bill of fare for diabetics may be constructed quite liberal enough to satisfy the palate of most reasonable persons by whom it is attainable.

FOOD AND DRINK ADMISSIBLE.—Shell-fish.—Oysters and clams, raw and cooked in any way, without the addition of flour.

Fish of all kinds, fresh or salted, including lobsters, crabs, sardines, and other fish in oil.

Meats of every variety except livers, including beef, mutton, chipped dried beef, tripe, ham, tongue, bacon, and sausages; also poultry and game of all kinds, with which, however, sweetened jellies and sauces should not be used.

Soup.—All made without flour, rice, vermicelli, or other starchy substances, or without the vegetables named below as inadmissible. Animal soups not thickened with flour, beef-tea, and broths.

Vegetables.—Cabbage, cauliflower, brussels-sprouts, broccoli, green string-beans, the green ends of asparagus, spinach, dandelion, mushrooms, lettuce, endive, coldslaw, olives, cucumbers fresh or pickled, radishes, young onions, water-cresses, mustard and cress, turnip-tops, celery-tops, or any other green vegetables.

Fruits.—Cranberries, plums, cherries, gooseberries, red currants, strawberries, apples, without sugar. Or they may be stewed with the addition of bicarbonate of sodium instead of sugar. (See below.)

Bread and cakes made of gluten, bran, or almond flour, or inulin, with or without eggs and butter. Griddle-cakes, pancakes, biscuit, porridges, etc. made of these flours. Where especial stringency is required these should be altogether omitted.

Eggs in any quantity and prepared in all possible ways, without sugar or ordinary flours.

Nuts.—All except chestnuts, including almonds, walnuts, Brazil-nuts, hazel-nuts, filberts, pecan-nuts, butternuts, cocoanuts.

Condiments.—Salt, vinegar, and pepper in moderate quantities.

Jellies.—None except those unsweetened. They may be made of calf's-foot or gelatin and flavored with wine.

Drinks.—Coffee, tea, and cocoa-nibs, with milk or cream, but without sugar; also milk, cream, soda- (carbonated) water, and all mineral waters freely; acid wines, including claret, Rhine, and still Moselle wines, very dry sherry; unsweetened brandy, whiskey, and gin. No malt liquors, except those ales and beers which have been long bottled, and in which the sugar has all been converted into carbonic acid and alcohol.

Vegetables to be especially Avoided.—Potatoes, white and sweet, rice, beets, carrots, turnips, parsnips, peas, and beans; all vegetables containing starch or sugar in any quantity.

The following list, including essentially the same articles, but arranged in the shape of a true bill of fare, by Austin Flint, Jr.,⁴⁹ will be found very convenient:

BILL OF FARE FOR DIABETES.—Breakfast.—Oysters stewed, without flour; clams stewed, without flour. Beefsteak, beefsteak with fried onions, broiled chicken, mutton or lamb chops; kidneys, broiled, stewed, or devilled; tripe, pigs' feet, game, ham, bacon, devilled turkey or chicken, sausage, corned-beef hash without potato, minced beef, turkey, chicken, or game with poached eggs. All kinds of fish, fish-roe, fish-balls, without potato. Eggs cooked in any way except with flour or sugar, scrambled eggs with chipped smoked beef, picked salt codfish with eggs, omelets plain or with ham, with smoked beef, kidneys, asparagus-points, fine herbs, parsley, truffles, or mushrooms. Radishes, cucumbers, water-cresses, butter, pot-cheese. Tea or coffee, with a little cream and no sugar. (Glycerin may be used instead of sugar if desired.) Light red wine for those who are in the habit of taking wine at breakfast.

Lunch or Tea.—Oysters or clams cooked in any way except with flour; chicken, lobster, or any kind of salad except potato; fish of all kinds; chops, steaks, ham, tongue, eggs, crabs, or any kind of meat; head-cheese. Red wine, dry sherry, or Bass's ale.

Dinner.—Raw oysters, raw clams.

Soups.—Consommé of beef, of veal, of chicken, or of turtle; consommé with asparagus-points; consommé with okra, ox-tail, turtle, terrapin, oyster, or clam, without flour; chowder, without potatoes, mock turtle, mullagatawny, tomato, gumbo filet.

Fish, etc.—All kinds of fish, lobsters, oysters, clams, terrapin, shrimps, crawfish, hard-shell crabs, soft-shell crabs, (No sauces containing flour.)

Relishes.—Pickles, radishes, celery, sardines, anchovies, olives.

Meats.—All kinds of meat cooked in any way except with flour; all kinds of poultry without dressings containing bread or flour; calf's head, kidneys, sweetbreads, lamb-fries, ham, tongue; all kinds of game; veal, fowl, sweetbreads, etc., with curry, but not thickened with flour. (No liver.)

Vegetables.—Truffles, lettuce, romaine, chicory, endive, cucumbers, spinach, sorrel, beet-tops, cauliflower, cabbage, brussels-sprouts, dandelions, tomatoes, radishes, oyster-plant, celery, onions, string-beans, water-cresses, asparagus, artichoke, Jerusalem artichokes, parsley, mushrooms, all kinds of herbs.

Substitutes for Sweets.—Peaches preserved in brandy without sugar; wine-jelly without sugar, gelée au kirsch without sugar, omelette au rhum without sugar; omelette à la vanille without sugar; gelée au rhum without sugar; gelée au café without sugar.

Miscellaneous.—Butter, cheese of all kinds, eggs cooked in all ways except with flour or sugar, sauces without sugar or flour. Almonds, hazel-nuts, walnuts, cocoanuts. Tea or coffee with a little cream and without sugar. (Glycerin may be used instead of sugar if desired.) Moderately palatable ice-creams and wine-jellies may be made, sweetened with pure glycerin; but although these may be quite satisfactory for a time, they soon become distasteful.

Alcoholic Beverages.—Claret, burgundy, dry sherry, Bass's ale or bitter beer. (No sweet wines.)

Prohibited.—Ordinary bread; cake, etc. made with flour or sugar; desserts made with flour or sugar; vegetables, except those mentioned above; sweet fruits.

⁴⁹ "On the Treatment of Diabetes Mellitus," a paper read before the American Medical Association at its meeting in Washington, May, 1884, and published in the Journal of the association July 12, 1884. I have so far modified the bill of fare as to permit the use of milk, which Flint excludes.

One of the foods the omission of which is most illy borne by the diabetic, however great his previous indifference to it, is wheaten bread, while the substitutes which have been at different times suggested for it very imperfectly supply its place. Perhaps the best known of these is the bread made of gluten flour. It was suggested by Bouchardat in 1841, and is made by washing the ordinary wheat flour to free it from starch.⁵⁰

⁵⁰ The Health Food Company, of 74 Fourth Avenue, N.Y., prepare a gluten flour by first removing the five bran-coats, pulverizing the cleaned berry by the cold-blast process, stirring the powder into iced water, and precipitating the gluten, cellulose, and mineral matters, siphoning off the water holding in suspension the starch, and drying out the precipitate. In this manner the salts of the wheat are retained. A purified gluten made by the Health Food Company is deprived of the cellulose walls of the cells in which the gluten granules are held. Directions for making gluten bread and cakes of various kinds are furnished by the company on application.

Gluten flour, however prepared, contains some starch, as indeed it must if bread is to be made out of it; and I confess to having been a good deal disappointed in its use. I have known the sugar absent in a selected diet to return when gluten bread was permitted, and again disappear on its withdrawal. Of course gluten flour contains less starch than the ordinary wheat flour, and there may be cases where the starch in the former can be assimilated when the quantity in the latter cannot be. The gluten may be made into porridge.⁵¹

⁵¹ Gluten porridge is made by stirring the gluten into boiling water until thick enough, and then keeping up the boiling process for fifteen minutes. A little salt and butter are added at the close to improve the flavor, and it may be eaten with milk or cream.

A method of getting rid of the starch and sugar in bread, suggested by Liebig and tried by Vogel, consists in converting the starch into sugar by the action of diastase and dissolving out the sugar thus produced. This is accomplished by treating thin slices of bread with an infusion of malt. The bread is then washed, dried, and slightly toasted.

Another substitute for wheaten flour is the bran flour whence the starch is removed by washing.⁵² The bran itself, according to Parkes,⁵³ sometimes contains as much as 15 per cent. of nitrogenous matter, 3.5 per cent. of fats, and 5.7 per cent. of salts. It is therefore not wholly innutritious, although the salts are washed out in removing the starch. It is considered especially useful when there is constipation, the slightly irritant properties of the bran aiding in maintaining a proper peristalsis and action of the bowels. These irritant properties are, however, inversely as the degree of comminution. The bran flour may be made with milk and eggs into a variety of cakes, of which the best known are those made according to Camplin's directions.⁵⁴

⁵² A very carefully prepared bran flour, as well as a wheat-gluten flour, is prepared by John W. Sheddon, pharmacist, corner of Broadway and Thirty-fourth street, New York City.

⁵³ Practical Hygiene, 5th ed., Philadelphia, 1878, p. 222.

⁵⁴ The following are Camplin's directions for making biscuit of bran flour: To one quarter of a pound of flour add three or four fresh eggs, one and a half ounces of butter, and half a pint of milk; mix the eggs with a little of the milk, and warm the butter with the other portion; then stir the whole together well; add a little nutmeg or ginger or other agreeable flavoring, and bake in small forms or patterns. The cake, when baked, should be about the thickness of an ordinary captain's biscuit. The pans must be well buttered. Bake in rather a quick oven for half an hour. These cakes or biscuits may be eaten by the diabetic with meat or cheese for breakfast, dinner, or supper; at tea they require rather a free allowance of butter, or they may be eaten with curd or any soft cheese.

Where extreme restriction of diet is not required the ordinary bran bread of the bakers may be used. The unbolted flour of which this is made of course contains the starchy principles, but in consequence of the retention of the bran the proportion of starch is less. The cold-blast flour of the Health Food Company is said to contain the nutritious, but not the innutritious, parts of the bran.⁵⁵

⁵⁵ It is made by pulverizing the carefully cleaned wheat by a compressed, cold air blast, which strikes the wheat and dashes it to atoms.

The almond food suggested by Pavy is another substitute for bread. The almond is composed of 54 per cent. of oil, 24 per cent. of nitrogenized matter known as emulsin, 6 per cent. of sugar, and 3 per cent. of gum, but no starch enters into its composition. Theoretically, therefore, the food should be everything that can be desired if the gum and sugar can be removed. The latter is done by treating the powdered almonds with boiling water slightly acidulated with tartaric acid, or soaking the almonds in a boiling acidulated liquid which may form a part of the process for blanching. The boiling and acid are necessary to precipitate the emulsin, which would otherwise emulsify the oil of the almond. Pavy speaks well of biscuit made of almond flour and eggs, which he says go very well with a little sherry or other wine, although he admits they are found too rich by some for ordinary consumption. One person only under my observation has used the almond food, and found it unpalatable.

Seegen recommends an almond food made as follows: Beat a quarter of a pound of blanched sweet almonds in a stone mortar for about three-quarters of an hour, making the flour as fine as possible; put the flour thus obtained into a linen bag, which is then immersed for an hour and a quarter in boiling water acidulated with a few drops of vinegar. The mass is thoroughly mixed with three ounces of butter and two eggs; the yolks of three eggs and a little salt are added, and the whole is to be stirred briskly for a long time. A fine froth made by beating the white of the three eggs is added. The whole paste is now put into a form smeared with melted butter and baked by a gentle fire.

Biscuits made of inulin, the starchy principle largely contained in Iceland moss, were suggested by Kuelz. Although a starch, it is one of the assimilable ones alluded to, of which small quantities at least may be taken as food without appearing in the urine as sugar. The biscuits are made with the addition of milk, eggs, and salt, and are inexpensive.

To some persons sugar is almost as imperative a necessity as bread, although to many it is not a very great sacrifice to omit it from ordinary cooking, if not from tea and coffee. For the latter it is just as well to dispense with sugar altogether. But where patients feel that they must have some substitute for sugar, glycerin has been suggested for this purpose, at least for sweetening tea and coffee. But Pavy has noted⁵⁶ that under the use of glycerin the urine increased from three and three and three-fourth pints to between five and six pints, and the sugar from 1100 grains to 3000 grains per diem, in the course of three days. Its withdrawal was followed by a prompt fall in both the urine and sugar, a return to it by a second increase, and subsequent withdrawal by another decline. Along with the increase of urine and sugar came also more thirst and discomfort. An examination of the chemical composition of glycerin would seem to confirm these results of experience. Glycerin is represented by C₃H₈O₃, sugar by C₆H₁₂O₃, and glycogen by C₆H₁₀O₅; whence it is evident that a conversion of glycerin into sugar may take place in the liver. These facts seem to show conclusively that glycerin is no suitable substitute for sugar. I therefore do not use it.

⁵⁶ On Diabetes, London, 1869, p. 259.

From what has been said it may be inferred that sugar of milk, mannite, and lævulose, or fruit-sugar, are admissible where sugar is demanded. They may be tried, but the urine should be carefully examined under their use, and if glycosuria occur or be increased they should be promptly omitted.

Almost every purpose of sugar in the cooking of acid vegetables is served by bicarbonate of sodium or potassium. As much bicarbonate of potassium to the pound as will lie upon a quarter of a dollar will neutralize the acidity of most fruits which require a large amount of sugar to mask this property. In this manner cranberries, plums, cherries, gooseberries, red currants, strawberries, apples, peaches, and indeed all fruits to which sugar is usually added in the cooking, become available to the diabetic.

In the matter of drinks, where the patient is not on a skim-milk diet, which usually affords as much liquid as is required by the economy, little restraint need be placed upon the consumption of water, which is demanded to replace that secreted with the sugar. Instead of water, Apollinaris water, Vichy, or the ordinary carbonated water may be used if preferred, and to many they are much more refreshing by reason of the carbonic acid they hold in suspension. Apollinaris water is particularly so, and one of my patients, who recovered completely under a suitable selected diet with which this mineral water was permitted, insists that it was that which cured her.

Where a simple selected diet is adopted, tea and coffee without sugar are usually permitted. The propriety of the substitutes for sugar already referred to must be determined by circumstances.

Of distilled and fermented liquors, moderate quantities of whiskey and brandy, dry sherry and madeira, the acid German and French wines—in fact, any non-saccharine wines—may be permitted. A medical friend who reports himself about cured of diabetes writes me that he has consumed eighty gallons of Rhine wine since he began to adhere closely to a diabetic diet. On the other hand, the free use of the stronger alcoholic drinks has been charged with causing diabetes, and I have known such use to produce a recurrence of sugar. No malt liquors, except those in which the sugar has been completely converted into carbonic acid and alcohol, should be used. Bass's ale may be allowed where no especial stringency is required.

HYGIENIC TREATMENT.—The patient should be surrounded by the most favorable hygienic influences. He should sleep in well-ventilated rooms; pass much time in the open air; bathe regularly, but not in water that is very cold, and especially the body should not be long submerged in cold water, as the liver must share the general internal hyperæmia incident to prolonged cooling of the skin, and increased glycosuria may result. I have known sugar to reappear after a prolonged drenching of the skin of patients overtaken by a rainstorm. Perhaps the most suitable time for the hot or tepid bath is on retiring in winter, but in summer it may be taken on rising. Thorough friction of the entire body should be practised after the bath or independently of it. An ounce or two of sodium carbonate may be added to it with advantage, as it softens the skin and facilitates the removal of the effete epithelium. The bowels should be kept regularly open, as the effect of their confinement is to produce torpor and congestion of the liver.

Certain natural mineral waters have always enjoyed a reputation for the cure of diabetes, and notably those of Vichy and Carlsbad. The former is an alkaline water with a slight laxative tendency, and the latter a decided aperient alkaline-saline water; and it is not unlikely that they owe a part of their good effects to an action upon the liver and upper bowel. This seems the more likely because Carlsbad, which enjoys the highest reputation, contains a far larger proportion of chlorides and sulphates, which are purgative. Vichy water contains 35 grains of carbonates to the pint, and Carlsbad 11, but the latter contains twice the proportion of chlorides, or 8 grains to the pint, and ten times as much sodium sulphate, or 19 grains to the pint. They may be used as adjuvants to the treatment, a pint of Vichy or half as much Carlsbad in the morning. Being imported waters, they are comparatively expensive, and I know of no American waters which closely approach them in composition.

Of American waters, the Saratoga Vichy contains twice as much chlorides as the Carlsbad, 17.7 grains to the pint, but no sulphates. It contains about the same amount of carbonates as Vichy. It is therefore a saline-alkaline water, and may be expected to serve the purposes of Vichy and some of those of Carlsbad, for which it may be substituted. Most of the American mineral waters vaunted as useful in diabetes will be found, on comparison with these waters, to be chemically indifferent, and therefore about as useful as so much ordinary spring-water. Of the Crab Orchard Springs in Kentucky, the Sowder's spring contains 25 grains of sulphate of sodium and magnesium and 7 grains of sodium chloride to the pint, therefore about the same proportion of the two substances combined as Carlsbad; yet I am not aware that these waters have any reputation in diabetes. The waters of Bedford Springs, Pennsylvania, also approximate them in the proportion of sulphates of sodium and magnesium.

Other Saratoga waters have an undoubted action on the liver through their chlorides, and may be used in lieu of the European waters above referred to, and of the Saratoga Vichy, when these cannot be obtained; such are the Geyser spring, which contains 70 grains of chlorides to the pint, and the Hathorn, containing 63 grains.

MEDICINAL TREATMENT.—While the dietetic treatment, and especially the skim-milk treatment, of diabetes mellitus is much to be preferred for its results over an exclusively medicinal treatment, and is of itself sufficient to control, if not to cure, a large number of cases, yet instances arise in which it is insufficient to complete the removal of sugar from the urine, and there are others in which it is impossible for various causes to carry out such treatment.

In my book on Bright's Disease and Diabetes, published three years ago, I gave the preference of drugs to ergot; but since then extended opportunities have convinced me that codeia is a far more efficient remedy. Repeated comparative trials of this drug in the wards of the Philadelphia Hospital and elsewhere have satisfied me of this. The trials have been made while the patients were upon a mixed diet, which I hold to be the only fair way of arriving at a knowledge of the true value of a drug in the disease. Codeia was first suggested by Pavy in lieu of opium and morphia, which had long been used, his reason being that it did not produce the same narcotic effect. Favorable reports upon its use have been made by Foster, Image, Brunton, R. Shingleton Smith, Cavafy, Austin Flint, Sr., Harvey L. Byrd, and others. It may be given in pill or solution. One should begin with ¼ of a grain three times a day, increasing ¼ of a grain daily until the sugar disappears or the remedy ceases to have any effect, or until drowsiness is produced. Thus gradually increasing, I have reached as high as 47 grains in a day. Cavafy has given 15 grains three times daily.

Opium—which is said to have been used by Aetius for this disease—or morphia might be used if codeia cannot be obtained, but they are less efficient, more dangerous, and more apt to produce the troublesome symptom of constipation. MacGregor⁵⁷ gave in one case 60 grains of opium, and in another 90 grains, in the twenty-four hours.

⁵⁷ London Medical Gazette, 1837.

While I have seen the most striking results upon the quantity both of sugar and urine during the administration of codeia, and at the same time have noted a gain in flesh and strength, I cannot say that I have ever seen a case totally recover under its use. Such cases are, however, reported by others. I have always used it in the very worst cases, where dietetic measures had also failed to remove the sugar. As to the mode of action of codeine, we can only speculate. It may be said that it quiets the irritation of the vaso-motor centre, whence result the glycosuria and other symptoms of diabetes.

Next to codeine in efficiency, of drugs, is ergot. The favorable results of its use are more easily explained by its physiological action—contraction upon the muscular walls of blood-vessels—than those of codeine, but it is not so efficient a remedy. It may be used by beginning with half a drachm, and increasing to a drachm, four times a day. Larger doses than this, as much as half an ounce four times a day, have been given, but the stomach rarely permits their continuation for any length of time.

Bromide of potassium, an old remedy for diabetes, has recently been revived and much lauded by the French physicians, but I have never found any results from its use. I can understand, however, how in certain cases of nervous origin it may be useful.

Comparatively recently, Clemens of Frankfort-on-the-Main has recommended the use of what he terms brom-arsen or bromide of arsenic. The dose is one-forty-eighth of a grain three times a day, gradually increased by this same amount until one-sixth or one-fifth of a grain is given daily.⁵⁸ Clemens, however, unites with its use a dietetic treatment. I have used it in connection with an unselected diet, and have not found the results claimed by Clemens. It is, however, both tonic and sedative, and as such is to be recommended in conjunction with other measures.

⁵⁸ The late Mr. R. F. Fairthorne, with Mr. James T. Shinn, apothecary, cor. Broad and Spruce streets, Philadelphia, prepared for me a solution of bromide of arsenic in the following manner: 77 grains of metallic arsenic in powder are added in small portions to 240 grains of bromine, the latter being placed in a long test-tube immersed in ice-water to control the otherwise violent reaction. One hundred grains of the tribromide thus obtained are dissolved in sufficient distilled water to make ten fluidounces. One minim will then contain one-forty-eighth of a grain.

Arsenic itself has some reputation in the treatment of diabetes, based upon the observation of Salkowsky that glycogen diminishes in the livers of animals poisoned with arsenic. It is at least a good adjuvant tonic. Leube gave it in diabetes in doses of one-third of a grain three times a day.

Strychnia is also very useful as a tonic, and may be used either alone or in the shape of the sulphate, or combined with arsenic and iron, or it may be given, perhaps preferably, in solution in combination with an acid. Given in combination with phosphoric acid, I believe it the most valuable tonic available in this disease.

To supply the phosphates, in which gluten bread is deficient, as well as for their tonic effect, the various preparations of phosphates are useful. The well-known compound syrup of the phosphates, or Parrish's chemical food, may be considered a type of these preparations. Every fluidrachm, which is a dose, contains 2½ grains of calcium phosphate, 1 grain of ferric phosphate, with fractions of a grain of sodium and potassium phosphate with free phosphoric acid. Similar is the solution of phosphates and phosphoric acid⁵⁹ known in this country as solution of phosphoric acid with iron, or the latter may be omitted.