QUANTITATIVE MICROSCOPIC DETERMINATIONS 149 



lines, in order that they may all be counted and yet none of them 

 counted twice. 



In the case of substances containing small starch granules as rice, 

 corn, chestnut and buckwheat, the 1/4000 cmm. areas of the hemacy- 

 tometer may be found more satisfactory. 



The suspension must be shaken thoroughly each time a hemacy- 

 tometer mount is to be made. 



(e) Making the Comparative Starch Count. Determine the exact 

 number of starch granules in 1 gram of the air dry pure substance 

 according to the method above outlined. This gives the exact number 

 of starch granules representing 1 per cent, of the substance under 

 examination and is the starch granule factor to be used in ascertainigg 

 the percentage of the starch or starch bearing substance present. Such 

 numerical starch granule factors must be very carefully determined . 

 Four or five different hemacytometer mountings should be carefully 

 counted and the average results obtained should be checked by making 

 additional counts of 2, 4 and perhaps 10 and even higher per cent, 

 suspensions of the substance. We will suppose that a pancake flour 

 shows the following starch counts according to the proposed method : 



Corn starch, 226,000,000 granules per gram. 



Wheat starch, 665,000,000 per gram. 



Let us suppose that corn meal contains 1,305,000,000 starch 

 granules per gram and wheat flour contains 910,000,000 starch gran- 

 ules per gram, then we would find from the above figures that the 

 pancake flour under examination was composed of corn meal 25 per 

 cent, and wheat flour 75 per cent. 



(/ ) Estimating the Percentage of Starch in the Dextrinized or Otherwise 

 Modified State. The method is applicable provided the identity of the 

 starch is still recognizable. The number of modified undissolved 

 dextrin granules still recognizable as starch granules will indicate the 

 percentage of air dry starch which it represents. 



(g) Starch Percentage Tables. It is desirable to prepare accurate 

 tables giving the average number of starch granules in 1 gram of the 

 starch, dextrin, flour or other starch bearing substance, for the con- 

 venience of laboratory workers. Such tables must be prepared from 

 counts made of meals, flours, seeds and grains, reduced to a very fine 

 powder (meal). 



In making comparative starch determinations as explained under 

 (e), all vegetable substances of a cellular nature must be reduced to a 

 very fine powder in order to make it possible to count all of the starch 

 granules representing 1 gram of the substance under examination. 



Until numerical starch percentage tables, as above suggested, are 



