28 



METHODS USED IN THE STUDY OF STARCHES. 



figures beyond the time-limit of complete gelatinization, 

 and that the figures have no value for comparison in cases 

 of starches which likewise are very quickly gelatinized, 

 unless by averages obtained from frequently repeated 

 experiments. 



When gelatinization occurs very slowly it often is 

 easier, after having made the count in the field, to deter- 

 mine the number of grains gelatinized and partially 

 gelatinized, as for instance when only 1 per cent of the 

 total starch is gelatinized at the end of 5 minutes or 5 

 or 10 per cent at the end of an hour. 



10. CONSTANCY OF RESULTS RECORDED BY THE FOKE- 

 QOINQ METHOD. 



It goes without saying that such experiments should 

 be carried out as far as possible under fixed conditions, 

 especially as regards the quantity of starch in relation 

 to the quantity of reagent. The variations in the quan- 

 tity of starch, in so far as constant results are concerned, 

 are absolutely negligible, as has been found not only in 

 the records of repeated experiments, but also in the 

 records of varieties of a species when the records should 

 be expected to be very close because of the starches being 

 nearly identical. The quantity of reagent used is in- 

 variably 2 drops, each reagent being kept in a 50 c.c. 

 bottle having a glass-stoppered finger pipette dropper 

 with a rubber tip. Under practically identical laboratory 

 conditions as regards quantity of starch, quantity of 

 reagent, temperature, and humidity the results recorded 

 by repeated experiments are either identical or vary 

 within limits that are so narrow as to be absolutely with- 

 out importance. Even marked variations in temperature 

 and humidity have not been found to be important, except 

 in rare instances. (See note under Amaryllis-Bruns- 

 vigia-Brunsdonnce, page 34.) 



Obviously, some variations, even though trifling, are 

 to be expected, so that in order to obtain constants a given 

 experiment should be repeated a sufficient number of 

 times and an average taken of the records, as in the 

 determination of melting-points. Experience has shown, 

 however, that in so far as the requirements of the present 

 exploratory research are concerned the results of a single 

 experiment carefully carried out are dependable within 

 narrow and wholly unimportant limits of error. The 

 chief sources of error to be guarded against are leakage 

 through the vaseline seal ; the presence of contaminating 

 substances in the starch ; certain peculiarities occasion- 

 ally observed in the behavior of starches towards certain 

 reagents; and errors in estimation when the reactions 

 are very rapid. Leakage through the vaseline seal is 

 sedulously to be avoided, and if a leak occurs the slide 

 and records must be discarded. 



The presence of oxalate crystals in the starch is by 

 no means uncommon, but no clear evidence has been 

 found to lead to the belief that, unless in exceptionally 

 large quantity, they in any way influence the course or 

 time of gelatinization by the reagents used. In the 



present research in Calanthe only were there even many 

 of these crystals; in the Phaius a few; arid none or 

 practically none in the other starches. Occasionally 

 foreign matter in the form of undetermined debris is 

 present which can not be gotten rid of by repeated wash- 

 ing, as in Tritonia pottsii. Such matter may affect the 

 polarization, iodine, and aniline reactions to a detectable 

 degree, but no effect has been noted in the other reac- 

 tions. With the exception of this starch all have been 

 free from such contamination. Erratic behavior of an 

 inexplicable character has upon rare occasions been ob- 

 served in the use of the sulphide and salicylate solutions. 

 Finally, when the reactions are very rapid, while satis- 

 factory records may not be obtained for comparison with 

 those of other starches which gelatinize with similar 

 rapidity, changes in the concentrations of the reagents 

 can be made so as to lengthen the time of the reactions 

 and thus permit of satisfactory differentiation. 



Comparatively little importance is to be attached to 

 the polarization, iodine, gentian violet, and safranin 

 reactions when the reactions are close. Personal equa- 

 tion and incidental conditions are here not unimportant 

 factors that may greatly vary the limits of error of ex- 

 periment. In future investigations these agents might 

 with profit be discarded for better means of study unless 

 further experience brings out greater values than they 

 have thus far shown. 



11. REAGENTS USED IN QUALITATIVE INVESTIGATIONS. 



The methods used in this research are both quantita- 

 tive and qualitative, chiefly the former because of the 

 ease with which the data recorded can be reduced to 

 figures and charts. The qualitative reactions have been 

 studied especially by means of certain of the chemical 

 reagents that were selected from time to time because 

 of their especial adaptation to certain kinds of starches 

 to elicit qualitative phenomena, some reagents acting 

 better with some kinds of starches than with others. 

 Incidentally here and there special qualitative records 

 were made by the use of selenite, iodine, gentian violet, 

 safranin, and heat. In the qualitative reactions many 

 points of varying degrees of interest and importance were 

 brought out that can not be studied by the quantitative 

 methods described, some of equal or greater importance 

 than those obtained generally by the latter methods. 



In studying the starches of the Amaryllidaceae we 

 used chloral hydrate, nitric acid, potassium iodide, potas- 

 sium sulphocyanate, potassium sulphide, and sodium 

 salicylate, excepting in the Narcissi when the sodium 

 salicylate was omitted. Additional studies were occasion- 

 ally made with sodium hydroxide, sodium sulphide, co- 

 balt nitrate, copper nitrate, cupric chloride, barium chlo- 

 ride, or mercuric chloride. In studying the Lilliaceaa 

 we used chloral hydrate, chromic acid, potassium hydrox- 

 ide, cobalt nitrate, and cupric chloride ; in the Iridaceae, 

 chloral hydrate, hydrochloric acid, potassium iodide, 

 sodium hydroxide, and sodium salicylate; in Begonia, 

 chloral hydrate, chromic acid, pyrogallic acid, nitric acid, 



