181. Separation by Hydrolysis with Water at High Temperatures.—Instead of dissolving the starch with diastase, it may be brought into solution by heating with water under pressure. The former method employed of heating in sealed flasks has been entirely superceded by heating in an autoclave. The materials are best held in metal beakers furnished with a cover which prevents loss from boiling if the pressure should be removed too rapidly after the completion of the operation. The autoclave is a strong metal vessel capable of resisting the pressure of several atmospheres. It is furnished with a pressure gauge C and a safety valve D, as shown in the figure. The top is securely screwed on by means of a wrench, shown at the right hand side. In the figure a portion of the case is represented cut away to show the arrangement of the metal beakers inside.

In the method of Reinke, as practiced at the Halle station, and in this laboratory, about three grams of the starchy substance are placed in each of the metal beakers with twenty-five cubic centimeters of a one per cent lactic acid solution and thirty cubic centimeters of water. The contents of the beaker are thoroughly mixed and they are then heated for two and a half hours in the autoclave, at a pressure of three and a half atmospheres. The addition of the lactic acid is for the purpose of protecting any sugar which may be present from decomposition at the high pressure and temperature employed. After the completion of the heating, the autoclave is allowed to cool, the cover is removed and the beakers taken out and their contents washed with hot water into quarter liter flasks. After cooling, the volume is completed with cold water, and after standing for half an hour, with frequent shaking, the contents of the flasks are filtered and 200 cubic centimeters of the filtrate in each case converted into dextrose with hydrochloric acid in the usual way. In order to obtain agreeing results, it is highly necessary that the substance before treatment should be ground to a fine powder. The addition of the lactic acid, as practiced in the reinke method, tends to give somewhat high results, due probably to the hydrolytic action of the acid on the fiber present. When starchy bodies are heated in the autoclave for the determination of their starch by polarimetric methods, or for ordinary determinations, the use of lactic acid should be omitted.

Example.—The following data indicate the methods of calculation to be followed in the determination of the percentage of starch in the material by diastatic hydrolysis: Three grams of a barley were inverted by diastase, as directed above, the volume of the solution made a quarter of a liter, filtered, 200 cubic centimeters of the filtrate converted into dextrose by hydrochloric acid, the volume completed to half a liter with water and fifty cubic centimeters thereof oxidized by the alkaline copper solution in the usual way. The amount of copper obtained was 331 milligrams, corresponding to 174 milligrams of dextrose. The amount of malt extract used in hydrolyzing the barley mentioned above, was ten cubic centimeters. The diastatic solution inverted with hydrochloric acid and treated as indicated above, yielded 191 milligrams of copper, corresponding to ninety-eight milligrams of dextrose in ten cubic centimeters of the malt extract. The quantity of malt extract represented in the final determination of copper, however, was only one and six-tenths cubic centimeters. We then have:

Calculated on the proportion that dextrose is to starch, as ten is to nine, this is equivalent to 142 milligrams of starch. The percentage of starch in the original substance, therefore, was equivalent to 142 multiplied by 100, divided by 240, viz., 59.17.

182. Principles of the Methods of Determination.—In the approximately pure state in which starch exists in the trade, it may be determined by conversion into dextrose and estimating the latter by one of the methods given. It is probable that there is no known method by which starch can be entirely converted into dextrose, and all the methods of hydrolysis, when used for quantitive purposes, must be standardized, not by the theoretical quantity of dextrose which a given weight of pure starch should yield, but by the actual quantity obtained. Starch is not largely converted into dextrose by any of the diastatic ferments which produce principally maltose and dextrins. Recourse must therefore be had to strong acids. In practice, hydrochloric is the one usually employed. By the action of a hot mineral acid, not only is starch converted into dextrose, but also the dextrose found is subjected to changes. In such cases an opposing action seems to be exerted by the hydrolytic agent, a part of the dextrose formed suffering a partial condensation, and thus assuming a state of higher molecular weight, approaching the constitution of the dextrins. Another part of the dextrose may also suffer oxidation and thus disappear entirely in respect of the further steps in starch analysis.

In such cases, the best the analyst can do is to conduct the hydrolysis in as nearly as possible constant conditions, and to assume that the percentage of dextrose present at a given time bears a constant ratio to the quantity of starch hydrolyzed. In reality almost all the starch appears finally as dextrose, and by proceeding on the assumption noted above a fairly satisfactory accounting may be made of the remainder.

Starch being insoluble, it cannot be determined directly by its rotatory power. When heated for a few hours in contact with water at a high pressure, starch becomes soluble, and in this state has a fairly constant gyrodynat, viz., [a]_D = 197°.

Starch is also rendered soluble by rubbing it in a mortar for about ten minutes with an excess of strong hydrochloric acid, and in this way a quick approximate idea may be obtained of the percentage present. Starch prepared in this manner, however, has a strong reducing power on metallic salts, showing that a part of it has already, even in so short a time, assumed the state of maltose or dextrose. The gyrodynat of pure anhydrous starch in such conditions varies from [a]_D = 197° to [a]_D = 194°. Starch is also rendered soluble by boiling with salicylic acid, whereby a solution is obtained having a gyrodynat of [a]_D = 200°(circa). The methods of procedure for the analysis of starch will be set forth in detail in the following paragraphs.

183. Estimation of Water.—In prepared or commercial starches the water may be determined by heating in a partial vacuum. The temperature at first should be low, not exceeding 60°. After drying for an hour at that heat the temperature may be gradually increased. The last traces of water come off from starch with difficulty, and the final temperature may be carried a little beyond 100° without danger of decomposition.

Ost recommends the use of an atmosphere of hydrogen or illuminating gas.[148] One and a half grams of the finely powdered sample are placed in the drying tube described in paragraph 23, and heated in a stream of dry hydrogen. The temperature at first is kept at about 60° for several hours and is then gradually increased to 120°. Ost states that even at 150° the sample preserves its pure white color, but so high a temperature is not necessary. Maercker, at the Halle station, makes use of the same process, but employs illuminating gas instead of hydrogen. The importance of beginning the desiccation at a low temperature arises from the fact that at a higher temperature, before the greater part of the water is driven off, the starch will suffer a partial fusion and form a paste which is very difficult to dry. The dried sample must be kept in a stoppered vessel to prevent the absorption of hygroscopic moisture.

184. Estimation of Ash.—When the drying is accomplished in a flat platinum dish, the same sample may serve for incineration. Otherwise the incineration may be accomplished in another portion of the sample by following directions already given.[149]

185. Nitrogen.—Even very pure samples of starch may contain a little nitrogen which is most conveniently determined by moist combustion.[150]

As a rule, in commercial starches of good quality, the quantity of pure starch may be considered to be the remainder after subtracting the sum of the weights of water, ash and nitrogen multiplied by 6.25, from the original weight of the sample taken.

Samples of starch usually contain also traces of fat and fiber, and these when present in weighable quantities, should be determined and proper deductions made.

186. Hydrolysis with Acids.—The acids commonly chosen for hydrolyzing starch are sulfuric and hydrochloric. The former has the advantage of being more easily removed from the finished product but the latter performs the work with less damage to the sugars formed. For commercial purposes sulfuric and for analytical practice hydrochloric acids are commonly employed.

The best process for analytical purposes is the one proposed by Sachsse.[151] In this method the starch is heated with the hydrolyzing mixture in the proportion of three grams to 200 cubic centimeters of water and twenty of hydrochloric acid of 1.125 specific gravity, containing five and six-tenths grams of the pure gas. The heating is continued for three hours on a steam-bath. Maercker recommends, instead of the above procedure, heating for two hours at gentle ebullition in an oil-bath. In this method three grams of the starch are reduced to paste with 200 cubic centimeters of water, and then boiled for two hours with fifteen cubic centimeters of hydrochloric acid of 1.125 specific gravity. The erlenmeyers in which the hydrolysis takes place are heated in an oil-bath and are provided with reflux condensers made of long glass tubes on which some bulbs have been blown, as shown in the accompanying figure. In all cases after hydrolysis the solution is neutralized, made to a standard volume and an aliquot part, after filtration, diluted to contain an amount of dextrose suited to the use of the table by Allihn for calculating the percentage of sugar. In diluting the solution preparatory to the estimation of dextrose, it is well to remember that nine parts of starch will furnish theoretically ten parts of dextrose. Since three grams of the sample are used, containing approximately eighty-five per cent of starch, the quantity of dextrose present is a little less than three grams. The solution should therefore contain not less than 300 cubic centimeters.

187. Factor for Calculating Starch from the Dextrose Obtained.—If all the starch could be converted into dextrose without loss, the quantity of it could be easily calculated theoretically on the supposition that the formula of starch is (C₆H₁₀O₅)ₙ. The factor by this assumption is, starch = dextrose × 0.90. If the starch have the formula assigned to it by Nägeli, viz., C₃₆H₆₂O₃₁ the formula becomes, starch = dextrose × 0.918.

Ost prefers to work by Sachsse’s method and to use the factor 0.925 to convert the dextrose into starch.[152]

In view of all the facts in the case it appears that the analyst will reach nearly correct results by converting the starch into dextrose by heating for three hours at 100° with hydrochloric acid or for two hours at gentle ebullition as directed above, determining the resultant dextrose and multiplying the weight thereof by 0.92.

188. Polarization of Starch.—Starch may be prepared for polarization by dissolving it in cold hydrochloric acid. The process as carried out by Effront is as follows.[153] Five grams of starch are rubbed with twenty cubic centimeters of cold concentrated hydrochloric acid for nearly ten minutes or until the solution is quite clear. The volume is completed to 200 cubic centimeters with water and the solution polarized. By this process there is always produced a notable quantity of reducing sugars, and for this reason it must be admitted that a portion of the starch has suffered complete hydrolysis. Ost therefore recommends the use of an acid of 1.17 specific gravity, and the gyrodynat of the soluble starch thus produced is found to vary from [a]_D = 196°.3 to 196°.7. When acid of 1.20 specific gravity is employed the gyrodynat falls to [a]_D = 194.2.[154] For approximately correct work the solution with the weaker hydrochloric acid and subsequent polarization is to be recommended as the most rapid method for starch determination.

It will be of interest to add the observation that the gyrodynat of maltose has lately been redetermined by Ost, who finds it to be [a]_D²⁰ ° = 137°.04 ± 0.19.[155]

189. Solutions of Starch at High Pressure.—Starch may also be brought into a condition suited to polarization by dissolving in water at a high temperature and pressure. The solution is accomplished in an autoclave as described in 181.

From two to three grams of starch are used and from eighty to ninety cubic centimeters of water. The starch is first reduced to a pasty state by heating with the water and, when evenly distributed throughout the flask, is rendered soluble by heating from three to five hours in an autoclave at from two to three atmospheres. The material is entirely without action on an alkaline copper solution. After heating, the volume of the solution is completed to 100 cubic centimeters and it is then polarized. The gyrodynat of starch dissolved in this way varies from [a]_D = 196°.5 to 197°.[156]

Starch is prepared by Baudry for polarization by boiling with salicylic acid.[157] The gyrodynat of starch dissolved in this way is [a]_D = 200°.25.

190. Polarization after Solution in Dilute Nitric Acid.—Guichard recommends saccharification with ten per cent nitric acid (ten cubic centimeters strong acid, ninety cubic centimeters water).[158] This treatment, even after prolonged boiling, gives only a light straw color to the solution which does not interfere with its polarization with a laurent instrument.

In working on cereals four grams of the finely ground material, in which the bran and flour are intimately mixed, are used.

The material is placed in a flask of about 500 cubic centimeters capacity, with 100 cubic centimeters of the dilute acid. The flask is closed with a stopper carrying a reflux condenser. After boiling for an hour the contents of the flask are filtered and examined in the saccharimeter. The dextrose formed is determined by the polarimetric data and the quantity of starch transformed calculated from the dextrose. The following formula is used:

In this formula a = the rotation in angular degrees, v = the volume of the liquid and A = the starch transformed.

In this method no account is taken of the sucrose and other sugars which are present in cereals. In the case of sucrose the left-handed sugar produced by treatment with nitric acid would diminish the rotation to the right and thus introduce an error. On the other hand the dextrose formed from the fiber of the bran would be calculated as starch. If these two errors should be compensating the method might prove practical.

191. Rapid Estimation Of Starch.—For the rapid estimation of starch in cereals, cattle foods and brewery refuse, Hibbard recommends a method which is carried out as follows:

The malt extract is prepared by covering ground, dry malt with water containing from fifteen to twenty per cent of alcohol. The object of adding alcohol is to preserve the filtered extract. It exercises a slight retarding effect on the action of the diastase, but prevents the malt extract from fermenting. After standing for a few hours in contact with the malt, the liquid is separated by filtration and is then ready for use. The substance in which the starch is to be determined should be dry enough to be finely pulverized, but previous extraction with ether is omitted. Enough of the material to contain at least half a gram of starch is placed in a flask with fifty cubic centimeters of water and from one to two cubic centimeters of malt extract added. The mixture is at once heated to boiling with frequent shaking to prevent the formation of clots. The addition of the diastase before boiling is to aid in preventing the formation of lumps. After boiling a minute the mixture is cooled to 60° and from two to three cubic centimeters of the malt extract added. It is then slowly heated until it again boils, consuming about fifteen minutes, when, after cooling, it is tested with iodin for starch. If a blue color be produced the operation above described is repeated until it fails to reappear. The mixture is then made up to a standard volume, thrown on a linen filter and an aliquot part of the filtrate, representing from 200 to 300 milligrams of starch, is boiled with five cubic centimeters of hydrochloric acid, of thirty per cent strength, for half an hour. The total volume of the liquid before boiling should be completed to sixty cubic centimeters. By the method above described, it is claimed that the determination of starch in a cereal or similar substance can be completed within two hours. The chief amount of time saved is in the heating with the malt extract, which instead of being continued for two hours, as usually directed, can be accomplished in thirty minutes.[159]

192. Precipitation of Starch with Barium Hydroxid.—The tendency of carbohydrate bodies to unite with the earthy bases has been utilized by Asboth as a basis for the quantitive determination of starch.[160]

About three grams of the finely ground sample containing the starch, or one gram of pure starch, are rubbed up in a mortar with water and the detached starch remaining suspended in the wash water is poured off. This operation is repeated until all the starch is removed. In difficult cases hot water may be used. The starch thus separated is heated in a quarter liter flask to the boiling point to reduce it to the condition of paste. When the paste is cold it is treated with fifty cubic centimeters of the barium hydroxid solution, the flask closed and well shaken for two minutes. The volume is then completed to the mark with forty-five per cent alcohol, the flask well shaken and allowed to stand. In a short time the barium-starch compound separates and settles. Fifty cubic centimeters of the clear supernatant liquor are removed with a pipette, or the liquor may be passed through a filter and the quantity mentioned removed for titration of the residual barium hydroxid after the addition of a few drops of phenolphthalein solution.

The quantity of barium hydroxid remaining, deducted from the original quantity, gives the amount which has entered into composition with the starch; the composition of the molecule being BaOC₂₄H₄₀O₂₀, which contains 19.10 per cent of barium oxid and 80.90 per cent of starch.

The set solution of barium hydroxid must be preserved from contact with the carbon dioxid of the air. The burette should be directly attached to the bottle holding the set solution, by any of the usual appliances, and the air entering the bottle must be deprived of carbon dioxid. The water used in the work must be also free of air, and this is secured by boiling immediately before use.

Example.—A sample of flour selected for the analysis weighed 3.212 grams. The starch was separated and reduced to paste in the manner described above. Thirty and four-tenths cubic centimeters of tenth-normal hydrochloric acid were exactly neutralized by ten cubic centimeters of the barium hydroxid solution. After treatment as above described, fifty cubic centimeters of the clear liquor, corresponding to ten cubic centimeters of the added barium hydroxid, required 19.05 cubic centimeters of tenth-normal hydrochloric acid. Then 30.4 - 19.05 = 11.35, and 11.35 x 5 = 56.75, which number corresponds to the total titration of the residual barium hydroxid in terms of tenth-normal hydrochloric acid. This number multiplied by 0.0324, viz., starch corresponding to one equivalent of barium, gave 1.8387 grams of starch or 57.24 per cent of the weight of flour employed.

The barium hydroxid method has been given a thorough trial in this laboratory and the results have been unsatisfactory when applied to cereals. The principle of the process, however, appears to be sound, and with a proper variation of working details, it may become practical.

193. Disturbing Bodies in Starch Determinations.—Stone has made a comparison of the standard methods of starch determinations, and the results of his work show that in the case of pure starch all of the standard methods give approximately correct figures. For instance, in the case of a pure potato starch, the following data were obtained:

By inversion with hydrochloric acid, 85.75 per cent; by inversion with oxalic and nitric acids, 85.75 per cent; by solution in salicylic acid, 85.47 per cent; and by precipitation with barium hydroxid, 85.58 per cent.[161]

When these methods are used, however, for the determination of starch in its original state, the widest variations are secured. Stone shows that these variations are due chiefly to the inverting effect of the reagents employed upon the pentosans present. In experiments made with pure xylan obtained from wheat straw, the methods employed gave from 44.73 to 67.16 per cent of material, which would be calculated by the usual methods as starch. Stone also shows that the pentosans are practically unaffected by the action of diastase or malt extract. Pure xylan treated with diastase, under the condition in which starch is converted into maltose and other soluble carbohydrates, fails to give any subsequent reaction whatever with alkaline copper solution. In all cases, therefore, where starch occurs in conjunction with pentose bodies, it is necessary to separate it by diastatic action before applying any of the methods of conversion of the starch into dextrose or its precipitation by barium hydroxid.

194. Colorimetric Estimation of Starch.—The production of the intensely blue color which starch gives with iodin has been used not only as the basis of a qualitive method, but also of many attempts at quantitive determination. These attempts have, as a rule, been attended with very unsatisfactory results, due both to the extraordinary delicacy of the reaction and to the fact that starches of different origin do not always give exactly the same intensity of tint when present in the same quantity. At the present it must be admitted that little should be expected of any quantitive colorimetric test.

In case such a test is desired the procedure described by Dennstedt and Voigtländer may be followed.[162] A weighed quantity of the starch-holding material, containing approximately half a gram of starch, is placed in a two liter flask and boiled with a liter of water. After cooling, the volume is completed to two liters and the starch allowed to subside. Five cubic centimeters of the clear supernatant liquor are placed in a graduated cylinder holding 100, and marked in half cubic centimeters. One drop of a solution of iodin in potassium iodid is added and the volume completed to the mark. A half gram of pure starch is treated in the same way and different measured portions of the solution treated as above until the color of the first cylinder is matched. From the quantity of pure starch in the matched cylinder the quantity in the sample is determined. The test should be made in duplicate or triplicate. If a violet color be produced instead of a blue, it may be remedied by treating the sample with alcohol before the starch granules are dissolved.

195. Fixation of Iodin.—In addition to forming a distinctive blue color with iodin, starches have the power of fixing considerable quantities of that substance. The starches of the cereals have this power in a higher degree than those derived from potatoes. In presence of a large excess of iodin the starches of rice and wheat have a maximum iodin-fixing power of about nineteen per cent of their weight. When only enough of iodin is employed to enter into combination the percentage absorbed varies from nine to fifteen per cent. The absorption of iodin by starches is a matter of importance from a general chemical standpoint, but as at present determined has but little analytical value. It is evident, however, that this absorption must take place according to definite chemical quantities and the researches of investigators may in the future discover some definite quantitive method of measuring it.[163]

196. Identification of Starches of Different Origin.—It is often important, especially in cases of suspected adulteration, to determine the origin of the starch granules. For this purpose the microscope is the sole resort. In many cases it is easy to determine the origin of the starch by the size or the shape and marking of the grains. In mixtures of more than one kind of starch the distinguishing features of the several starches can be clearly made out in most instances. There are, however, many instances where it is impossible to discriminate by reason of the fact that the characteristics of starch granules vary even in the same substance and from year to year with varying conditions of culture.

In many cases the illustrations of the forms and characteristics of starch granules which are found in books are misleading and no reliance can be placed on any illustrations which are not either photographs or drawings made directly from them. In the microscopic study of starches the analyst will be greatly helped by the following descriptions of the characteristic appearance of the granules and the classifications based thereon.[164]

197. Vogel’s Table of the Different Starches and Arrowroots of Commerce.—A. Granules simple, bounded by rounded surfaces.

• I. Nucleus central, layers concentric.
• a. Mostly round, or from the side, lens-shaped.
• 1. Large granules 0.0396-0.0528 mm, rye starch:
• 2. Large granules 0.0352-0.0396 mm, wheat starch:
• 3. Large granules 0.0264 mm, barley starch.
• b. Egg-shaped, oval, kidney-shaped: Hilum often long and ragged:
• 1. Large granules 0.032-0.079 mm, leguminous starches.
• II. Nucleus eccentric, layers plainly eccentric or meniscus-shaped.
• a. Granules not at all or only slightly flattened:
• 1. Nucleus mostly at the smaller end; 0.06-0.10 mm, potato starch:
• 2. Nucleus mostly at the broader end or towards the middle
• in simple granules; 0.022-0.060 mm, maranta starch.
• b. Granules more or less strongly flattened.
• 1. Many drawn out to a short point at one end.
• a. At most 0.060 mm long, curcuma starch:
• b. As much as 0.132 mm long, canna starch:
• 2. Many lengthened to bean-shaped, disk-shaped, or flattened;
• nucleus near the broader end; 0.044-0.075 mm, banana starch:
• 3. Many strongly kidney-shaped; nucleus near the edge;
• 0.048-0.056 mm, sisyrinchium starch:
• 4. Egg-shaped; at one end reduced to a wedge, at the other enlarged;
• nucleus at smaller end; 0.05-0.07 mm, yam starch:

B. Granules simple or compound, single granules or parts of granules, either bounded entirely by plain surfaces, many-angled, or by partly round surfaces.

• I. Granules entirely angular.
• 1. Many with prominent nucleus: At most 0.0066 mm, rice starch:
• 2. Without a nucleus: The largest 0.0088 mm, millet starch:
• II. Among the many-angled also rounded forms.
• a. No drum-shaped forms present, angular form predominating.
• 1. Without nucleus or depression very small;
• 0.0044 mm, oat starch:
• 2. With nucleus or depression; 0.0132-0.0220 mm.
• a. Nucleus or its depression considerably rounded;
• here and there the granules united into differently
• formed groups; buckwheat starch:
• b. Nucleus mostly radiate or star-shaped; all the
• granules free; maize (corn) starch:
• b. More or less numerous kettledrum and sugar-loaf like forms.
• 1. Very numerous eccentric layers; the largest granules
• 0.022-0.0352 mm, batata (sweet potato) starch:
• 2. Without layers or rings; 0.008-0.022 mm.
• a. In the kettledrum-shaped granules the nucleal
• depression mostly widened on the flattened
• side; 0.008-0.022 mm, cassava starch:
• b. Depression wanting or not enlarged.
• aa. Nucleus small, eccentric; 0.008-0.016 mm,
• pachyrhizus starch:
• bb. Nucleus small, central, or wanting.
• aaa. Many irregular angular forms;
• 0.008-0.0176 mm, sechium starch:
• bbb. But few angular forms; some with radiate,
• nucleal fissure; 0.008-0.0176 mm, chestnut starch.

C. Granules simple and compound; predominant forms, oval, with eccentric nucleus and numerous layers; the compound granule made up of a large granule and one or more relatively small kettledrum-shaped ones; 0.025-0.066 mm, sago starch.

198. Muter’s Table for the Detection of Starches
when Magnified about 230 Diameters.

[All measurements are given in decimals of an inch.]

Group I: All more or less oval in shape and having both hilum and rings visible.

Group II: With strongly developed hilum more or less stellate.

Group III: Hilum and rings practically invisible.

Group IV: More or less truncated at one end.

Group V: All granules more or less polygonal.

199. Blyth’s Classification.—Blyth gives the following scheme for the identification of starch granules by their microscopic appearance.[165]

Division I.—Starches showing a play of colors with polarized light and selenite plate:

The hilum and concentric rings are clearly visible, and all the starch granules, oval or ovate. Canna arrowroot, potato, arrowroot, calumba, orris root, ginger, galangal and turmeric belong to this division.

Division II.—Starches showing no iridescence, or scarcely any, when examined by polarized light and selenite:

Class I.—The concentric rings are all but invisible, and the hilum stellate. The bean, pea, maize, lentil, dari and nutmeg starches are in this class.

Class II.—Starches which have both the concentric rings and hilum invisible in the majority of granules: this important class includes wheat, barley, rye, chestnut, acorn, and many starches in medicinal plants.

Class III.—All the granules are truncated at one end. This class includes sago, tapioca and arum, several drugs and cinnamon and cassia.

Class IV.—In this class all the granules are angular in form and it includes oats, tacca, rice, pepper and ipecacuanha.

200. Preparation of Starches for Microscopical Examination.—The approximately pure starches of commerce may be prepared for microscopic examination by rubbing them up with water and mounting some of the suspended particles by one of the methods to be described below.

In grains, seeds and nuts the starch is separated by grinding with water and working through fine linen. The starch which is worked through is allowed to subside, again beaten up with water if necessary and the process continued until the grains are separated sufficiently for microscopic examination. A little potash or soda lye may be used, if necessary, to separate the granules from albuminous and other adhering matter. The analyst should have a collection of samples of all common starches of known origin for purposes of comparison.

The granules are mounted for examination by plain light in a medium of glycerol and camphor water. When polarized light is used the mounting should be in Canada balsam.[166] The reader can find excellent photomicrographs of the more common starches in Griffith’s book.[167]

201. Appearance in Balsam with Polarized Light.—Mounted in balsam the starches are scarcely visible under any form of illumination with ordinary light, the index of refraction of the granules and the balsam being so nearly alike. When, however, polarized light is used the effect is a striking one. It is very easy to distinguish all the characteristics, except the rings, the center of the cross being at the nucleus of the granule.

With the selenite plate a play of colors is produced, which is peculiar to some of the starches and forms the basis of Blyth’s classification.

202. Description Of Typical Starches.—The more commonly occurring starches are described by Richardson as they appear under the microscope magnified about 350 diameters.[168]

The illustrations, with the exception of the cassava starch, and the maize starch accompanying it were drawn by the late Dr. Geo. Marx from photographs made by Richardson in this laboratory. The two samples excepted were photographed for the author by Dr. G. L. Spencer.

Maranta Starch.—Of the same type as the potato starch are the various arrowroots, the only one of which commonly met with in this country being the Bermuda, the starch of the rhizome of Maranta arundinacea, and the starch of turmeric.

The granules are usually not so varied in size or shape as those of the potato, averaging about 0.07 millimeter in length as may be seen in Fig. 48. They are about the same size as the average of the potato, but are not often found with the same maximum or minimum magnitude, which circumstance, together with the fact that the end at which the nucleus appears is broader in the maranta and more pointed in the potato, enables one to distinguish the two starches without difficulty. With polarized light the results are similar to those seen with potato starch, and this is a ready means of distinguishing the two varieties, by displaying in a striking way the form of the granule and position of the hilum.

Potato Starch.—The starch grains of the potato are very variable in size, being found from 0.05 to 0.10 millimeter in length, and in shape from oval and allied forms to irregular and even round in the smallest. These variations are illustrated in Fig. 49, but the frequency of the smaller granules is not as evident as in some other cases. The layers are visible in some granules with great distinctness and in others hardly at all, being rather more prominent in the starch as obtained from a freshly cut surface. The rings are more distinct, too, near the hilum or nucleus, which in this, as in all tuberous starches, is eccentric, shading off toward the broader or more expanded portion of the granule. The hilum appears as a shadowy depression, and with polarized light its position is well marked by the junction of the arms of the cross. With polarized light and a selenite plate a beautiful play of colors is obtained. The smaller granules, which are nearly round, may readily be confused with other starches, but their presence serves at once to distinguish this from maranta or Bermuda arrowroot starch. Rarely compound granules are found composed of two or three single ones each with its own nucleus.

Ginger Starch.—This starch is of the same class as those from the potato and maranta and several others which are of underground origin. In outline the granules are not oval like those named, but more rectangular, having more obtuse angles in the larger ones and being cylindrical or circular in outline in the smaller, as indicated in Fig. 50. They average nearly the same size as maranta starch, but are much more variable, both in size and form. The rings are scarcely visible even with the most favorable illuminations.

Sago Starch.—This exists in two modifications in the market; as raw and as prepared sago. In the prepared condition it is characterized by a larger circular depression in the center of most of the granules. The rings are not visible. They are mostly circular in form or approaching it, and vary from 0.025 to 0.065 millimeter in diameter, as indicated in Fig. 51.