The introduction of such large amounts of water as have been sometimes found (up to fifty per cent) is said to be only made possible by warming the butter until it is quite soft and then stirring in water until it hardens sufficiently to retain it.

Excessive amounts of any preservatives other than salt, as saltpeter, borax or salicylic acid. By some, any addition of these substances is regarded as an adulterant.

Vegetable matters for bulk as starch from grain or potatoes, potatoes dried and ground, carrot scrapings, syrup, carregeen mucilage.

Animal products for bulk. White cheese, gelatine, butyric acid. Mineral matters for bulk and weight as chalk, soapstone, heavy spar gypsum, white lead, alum, water glass or sodium silicate and alkalies. The use of borax or alum is said to make the incorporation of larger proportions of water more easy. (Dietzsch, p. 206).

Fatty substances. Vegetable oils, as benne or sesame oil, cottonseed oil, mangosteen oil, rape-seed, cocoanut, palm, peanut, olive and other oils.

Animal fats, as beef, mutton and bone fat, lard and lard oil, etc.

Hurtful coloring matters as chrome yellow (chromate of lead) and coal tar colors, Victoria yellow (Saffron substitute). (Potassium Dinitrocresylate). Martius yellow (calcium or potassium Dinitronaphthylate), etc. Vegetable coloring matters ordinarily reckoned as harmless are not considered as adulterants by many. Such are annatto carrot juice, turmeric and marigold flowers. (Calendula off). The fruit of the winter cherry (Physalis alkekingi) has been mentioned, but Tollen's Handbuch off. Gesundheitswesens, vol. 1, p. 46, says that it is improbable.

DETERMINATION OF WATER AND OTHER NON-FATTY CONSTITUENTS. Determination of water and other non-fatty constituents.

It has been noted by Husson that when much water is present in a butter if a knife is plunged into it droplets of water appear in the track of the knife. He also notes that the butter is crumbly when large amounts of water are present.

In taking a sample for the quantitative determination of the water, care must be taken to obtain what fairly represents the average constitution of the sample. If the sample is soft it may sometimes be previously mixed with a stout spatula. If hard, a fair proportion of the interior and exterior must enter into the portion taken. Since the water and fat do not readily mingle the mixture of the two can seldom be made quite uniform. The simplest method in theory proposed for the determination of the water, consists in melting a known amount in a graduated tube, and reading off the volume of water. Although the volume of water is slightly increased by the curd and salt present in it, the increase is umimportant in this relation, and the relative proportion of the water by weight can be calculated. (Hassall). The difficulty with this method is that the curd only partially separates from the fat, and the line of demarcation between the water and fat is rendered indistinct.

A method used by Heeren, Lefeldt, Birnbaum and others consists in melting the butter as before mentioned and forcing a more complete separation in a short time by placing the graduated tube in an iron cylinder attached to a rotating arm. The apparatus described

and figured (first report, New York State Dairy Commissioner, p. 106), as in use for the determination of the volume per cent of cream in milk can be used, or a rougher device of the same character can be improvised. (v. Babo Landwirth correspblatt., Baden, 1863 p. 65. Martigny Die Milch, etc., 1, 1871-2, 195).

The use of some solvent for the fat as ether, or petroleum naphtha to facilitate the separation is a method adopted by some. Babo uses ether and the centrifugal apparatus already alluded to. Other chemists as Hoorn, Johanson, Caldwell (2d Report N. Y. State Board of Health, p. 526), and others use petroleum ether, and after thorough shaking allow the graduated tubes to stand, and read off.

An objection, though not a very serious one, to these methods is that the results are obtained in percentages by volume and not by weight, the usual form in which they are to be recorded or reported. On this account most analysts prefer the determination of water by weight.

Weighing the butter in a cylinder, adding about twice its weight of water, keeping the fat melted until clear, solidfying and cooling, and then running out the water and weighing the fat remaining is a method used by some. (Duflos and Hirsch, Die wichtigsten Lebensõeduerfnisse, etc). Griessmayer, Die Verfalschung der wichtigsten Nahrungs und Genussmittel, 1880. Elsner Die Praxis des Nahrungsmittel-Chemikers, 1880.

Dietzsch (Nahrungsmittel und Getraenke, 4th ed. Zurich, 1884, p. 214), accomplishes this in an ingenous manner by putting the mixture. of butter and warm water into a tube, corking it up, shaking and cooling in an inverted position so that the fat occupies the lower end of the tube and the water may be readily poured off.

These methods are properly a determination of the amount of fat and not of the water, though by them data can be obtained from which the approximate percentage of water may be calculated.

A practical objection to this method is that with many butters, as well as substitutes therefor, much of the curd holding moisture will remain suspended in the melted fat, rendering the complete separation of the water difficult or impossible. The water poured off contains the salt and a part only of the curd, so that with the most careful weighing it affords only an approximate estimation of the amount of fat present. Objections of this character are also presented by Cauldwell (Loc. cit).

Drying a weighed quantity of butter in a dish over a boiling water bath is another more accurate method (Blyth).

It requires some considerable time, during which there is a danger of some loss of the fat at the temperature required. The water being heavier sinks to the bottom of the melted fat and is protected by it against rapid evaporation. The addition of a small amount of absolute alcohol to cause the water and fat to commingle and allow water and alcohol to be driven out together, only partially remedies this difficulty since the curd acts as a sponge and obstinately retains the water. The method which has been found the most satisfactory (Angell and Hehner, p. 12), consists in drying a weighed portion of the sample in a capacious platinum dish over a low flame, the mass being continually stirred with a thermometer which should not indicate over 105 to 110 degrees C. (220 to 230 degrees F.), throughout (Blyth, foods).

The operation is finished when bubbles of steam cease to escape through the fat. The method requires constant attention, but is more rapid and satisfactory than drying over the water bath. From seven to ten grammes is a good quantity to use for this test. It is well to heat cautiously for some little time until the curd contracts in bulk from loss of water before allowing the thermometer bulb to come in contact with the curd, otherwise it attaches itself very firmly and is removed with difficulty. The small amount of fat adhering to the thermometer bulb may be rinsed into the dish with a few drops of ether, and the ether driven off by a few minutes heating on the water bath.

The other non-fatty adulterants mentioned, except coloring matters which usually remain with the fat, are frequently neglected by analysts because of their infrequency (Allen 2, p. 202). These analysts content themselves with a determination of the fat alone, or of the water and the fat. If the fat is eighty per cent or over they regard the sample as unadulterated so far as non-fatty constituents are concerned.

Bischoff (Handb. off. Gesundheitswesens, 2.509) states that in none of the samples of butter examined by him in Berlin, since 1875 (some 800), has he found any of the mineral or vegetable adulterations mentioned, though accusations of that sort have frequently been made. The term "bosh butter" has been applied according to Hassall (p. 430), to butters adulterated with starch or substances containing it. According to a witness before the New York Legislative Committee on public health (p. 227), the name was derived from the name of the town Hertogenbosch, in Holland where large amounts of artificial butter are made.

To determine the character of the non-fatty adulterants other than water, Hilger (Handb. der Hygien. Nahrungsmittel, Leipzig, 1882, p. 256), recommends taking five to ten grammes of the butter, adding to it in a test tube about twice as much water (preferably containing alcohol) and warming for some time at the melting point of the fat. The fatty layer retains the foreign coloring matters, the aqueous layer below dissolves the salt, borax, alum, alkali, water glass, salicylic acid, etc., while the insoluble portions, as starchy material, heavy spar, chalk, etc., settle to the bottom.

Dietzsch (Loc. cit.) gives a detailed method of testing the solution and sediment obtained by a similar treatment. The sediment is filtered out, then the caseine is dissolved off by ammonia, leaving starch and insoluble mineral matters to be tested by the microscope by iodine for starch and by other well known chemical methods for chalk, gypsum, heavy spar (barium sulphat), soapstone, etc.

In the water solution tests may be made for the soluble adulterants. J. B. Edwards, of Montreal, states that borax to the extent of one or two per cent is not, in his estimation, an adulteration. By a large number of chemists, however, the presence of any borax is regarded as a hurtful addition, since it is believed by them that even small amounts in food interfere materially with the processes of digestion, the effect being naturally more marked where the digestive powers are weak.

To detect lead, copper or zinc compounds, the two latter usually being present as the result of using brass, copper or zinc vessels in the manufacture or storing of butter Dietzsch (p. 217), also Tollens (Handb. off. Ges. p. 492), recommends boiling fifty grammes of the butter for

fifteen minutes with 100 to 120 cubic centimetres of water, and ten cubic centimetres pure hydrochloric or nitric acid, filtering through a wetted filter, and testing the acid solution for these metals with sulphuretted hydrogen, potassium ferrocyanide, etc.

The addition of gelatin to butter, presumably for the purpose of giving it more consistency in warm weather, is a matter which has come up but recently and has not yet been mentioned in the literature of the subject. The presence of the gelatin is very difficult of detection since in many of its properties it bears a close resemblance to the albuminoids of the curd. Besides this, our knowledge of gelatin and of the albuminoids in milk is as yet very incomplete.

A method for its detection was devised which seems to answer the requirements of the case tolerably well. The gelatin was extracted from fifty to seventy-five grammes of the butter by heating with water, the water solution partly cleared by filtration or straining through muslin, was then boiled with the addition of a drop of dilute acetic acid and a few drops of potassium ferrocyanide until the precipitate, which only forms after a time, turned blue and was in a state to filter readily. In the warm filtrate the gelatin was tested for by solution of tannic acid or chlorine gas. With most butters unadulterated with gelatin, only a slight turbidity was produced. With a few containing much buttermilk or with milk serum the turbidity was more dense, sometimes affording a flocculent precipitate. When gelatin was present a stringy precipitate of very different character and very decided was obtained. Examinations of the fat, so far as the experiments went, did not indicate that any reaction between the fat and the gelatin had taken place.

The addition has been chiefly suspected in the case of artificial butters, and with these the question comes up as to whether some gelatinoid substance may not be extracted from the membrane in the process of rendering the fat. The presence of some such substance is specified in some of the patents relating to the manufacture of oleomargarine and other fats. (Paris, Pat. No. 19,011, Jan. 1882, Huet. Eng. Pats. 5,249, 1881, and 134, 1882, Huet.)

Butyric acid is reported as an adulterant of butter by J. Zannı (Quelques essais sur des Beurres fondus, Constantinople, Fres. zeits. f. Anal. Chem. 23, 87). To detect it he recommends washing the butter fat and testing for and determining the butyric acid in the washings.

On account of the presumable infrequency of adulteration by organic non-fatty substances or mineral substances other than salt, it is a common practice among analysts to call all the mineral constituents of butter "salt," and the organic non-fatty substances "curd "caseine " (Allen 2, 204, etc).

The mineral constituents of butter, salt and mineral adulterants, are determined by incineration of weighed amounts of the material directly, or incinerating the residue after extracting the fat with ether or some similar solvent. Lead, if present, might be thus volatilized and escape detection, saltpetre also would be destroyed, but all other mineral adulterants enumerated will remain and can be detected by the well known methods of mineral analysis.

If a strong heat is applied before the removal of the water there is danger of mechanical loss by the boiling of the water beneath the layer of fat. After removing the water the fat can be burned off [Sen. Doc. No. 24.]

24

quietly. The salt often fuses about some of the partially burned carbon in the residue so as to protect it from oxidation, leaving the ash gray, or even black, and if the heat is continued until the ash is perfectly white, so much salt has been volatilized that the quantitative results are erroneous. If the ash is light gray, without long heating the results are sufficiently close. If black, the most accurate method is to dissolve all the soluble salts with water, and burn the carbonaceous residue, add the water solution, evaporate, ignite and weigh. Titration with standard silver nitrate solution will then show whether the mineral matter is entirely or only partially common salt. The fat may be determined approximately by the method of Dietzsch. The amount of fat is determined by melting the butter until the curd has settled, pouring off as much of the clear fat as possible into a weighed dish, dissolving off the remainder of the fat into the same dish with ether, petroleum spirit or carbon disulphide, evaporating off the solvent and weighing.

Some chemists use the solvent from the start without pouring off any portion of it. In the case of some butters where the curd separates with difficulty, this is the only satisfactory plan. The residue insoluble in the ether may be dried and weighed, giving curd and salt. If the amount of water is known the percentage of fat may be determined indirectly by difference, a plan recommended as accurate by Allen (2, p. 105), Hehner and Angell (p. 13, etc.).

Some prefer to mix the butter with sand or pounded glass and extract in a fat extraction apparatus as Soxhlets. Hilger (Handb. Hyg. Nahr. p. 256) Tollens, Soxhlet, Thom., etc.

The proportion of caseine or curd, or what is generally reckoned as such is usually obtained by difference. The approximate estimation of the fat by melting with water in a tube and pouring off the water has already been mentioned.

EXAMINATION OF THE FAT.

The most important point at the present day in the examination of butter is the determination of the character of the fat, and first, it may be stated that what is known in commerce or in daily life as a simple fat, consists almost invariably of a mixture of several different substances called "fats" by the chemists. In the chemical sense fat results from a combination of a fatty acid with glycerine, water being eliminated at the time of the union. Such fats are called glycerides, and the names by which they are designated terminate in in,"e. g., stearin is the glyceride of stearic acid, palmitine is the glyceride of palmitic acid, arachin, that of arachidic, etc., etc.

66

Glycerine has the property of uniting with one, two or three molecules of fatty acid affording mono-di or tri-glycerides designated according to the acid, e. g., mono-stearin, distearin, tristearin, monopalmatin, dipalmatin, tripalmatin, etc., etc.

In almost all the fats occurring in nature the fats are tri-glycerides, tristearin, tripalmatin, triolein, etc., and in speaking of natural products the prefix "tri" is omitted, being understood.

The mode of combination of glycerine with fatty acids may perhaps be best understood from the following example :