NOTES AND CONCLUSIONS. 



373 



and qualitative gelatinization reactions with a variety 

 of chemical reagents which represent a wide range of 

 differences in molecular composition. In some instances 

 the starch molecules alone or largely determine the 

 reaction, while in others both starch and reagent play 

 important parts, as in chemical reactions generally. 

 Thus, in the crystallographic studies of the hemo- 

 globin crystals and in the polarization react ions with 

 starch the molecules undergo no change ; hence the 

 reactions express peculiarities that are inherent to the 

 molecules. In other starch reactions, in the gentian- 

 violet and safranin reactions, the organization of the 

 molecules is either unaffected or affected to an unde- 

 tectable degree, the reactions being presumably adsorp- 

 tion phenomena; in the iodine reactions there is prob- 

 ably a feeble chemical combination of the iodine and 

 starch, but without apparent intermolecular disorgan- 

 ization ; in the temperature and chemical-reagent reac- 

 tions there is an intermolecular breaking down by a 

 process of hydration, with which process there may be 

 associated reactions that vary in character and number in 

 accordance with peculiarities in the composition of the 

 reagents. If the molecules of the starches from different 

 sources are in the form of stereoisomers, it follows, as a 

 corollary, that they must exhibit differences in their 

 behavior with different agents and reagents, and show 

 differences that are related to variation in the kind of 

 agent and in the composition and concentration of the 

 reagents. In other words, the reaction in each case is 

 conditioned by the kind of starch and the kind of 

 agent or reagent. 



RELIABILITY OF METHODS AS SHOWN BY CHARTS AND 

 CONFORMITY OF RESULTS COLLECTIVELY. 



It is obvious that tests of the reliability of the 

 methods employed in the differentiation of starches 

 from various sources are to be found in the agreement 

 of the results of repeated experiments and in the con- 

 formity of the results with established data of the syste- 

 matist. As stated in preceding paragraphs, the polari- 

 zation, iodine, and aniline methods are, notwithstanding 

 their crudity and limitations, reliable if the experiments 

 are carried out with sufficient care ; the temperature of 

 gelatinization method is accurate within very narrow 

 limits of error ; and the gelatinization method used in 

 the present research by means of chemical reagents is 

 practically exact. The first three methods are, owing 

 to their usually very restricted range of values, of very 

 much more usefulness in the differentiation of members 

 of a genus than of different genera, and this applies, 

 although to a less degree, to the temperature of gela- 

 tinization method ; while the chemical reagent method 

 has unlimited application to both intrageneric and in- 

 tergeneric differentiation, though the different rea- 

 gents have widely varying values. In comparing 

 these records with those of the systematist it is im- 

 portant to recognize that a slight change in molecular 

 constitution may give rise to very marked changes in 

 properties and that distinction must be made between 

 that which is definitely established and that which is ten- 

 tative in even the most advanced taxonomic system. All 

 things considered, it is remarkable how close in general 

 is the agreement of the data of these exceedingly dissimi- 

 lar methods of investigation. In fact, they are evidently 



mutually corrective and confirmatory, and where seeming 

 or actual disagreements exist it doubtless will be found 

 that further applications of the physico-chemical method 

 will demonstrate the reasons. 



Certain of the several forms of charts are of especial 

 value in showing the reliability of the methods used, 

 particularly those which are included in the groups D 1 

 to D GDI and E 1 to K 4(J. These charts have been given 

 somewhat detailed discussion in Sections 2 and 3 of 

 Chapter IV. Even a most cursory examination separ- 

 ately and together will demonstrate their taxonomic 

 values. In group D 1 to D GDI, in which are presented 

 the progress of gelatinization at definite time-intervals, 

 it is obvious from the characters of the curves, both in 

 courses in the individual charts and in the parent-hybrid 

 and the generic groups, that they are quite as dependable 

 as the data of the systematist. Were these records not 

 reliable, it seems clear that the curves would not take 

 regular but irregular or zigzag circumlinear courses, or 

 instead of being straight or practically straight lines be 

 irregular, etc. ; moreover, there would not be the con- 

 formity of the curves of the reactions with each reagent 

 that is found in each set of parent- and hybrid-stocks, 

 or in the sets belonging to each genus, excepting in the 

 latter when subgeneric divisions are represented. The 

 more or less marked subgeneric differences attest the 

 value of the method, and if in some instances they may 

 seem to be disproportionate to the differences of the sys- 

 tematist, this may be and doubtless is owing to a greater 

 sensitivity of the physico-chemical method. 



The plan adopted in the preparation of Charts E 1 to 

 E 46, in which composite curves of the reaction-intensi- 

 ties are exhibited, has proved in a very large measure 

 successful in eliciting varietal, species, subgeneric, and 

 generic peculiarities, but its essential defect is to be 

 found in the neglect of differences that were found dur- 

 ing the earlier periods of experiment. In the formula- 

 tion of these charts terminal data were used that is, the 

 time of complete or practically complete gelatinization 

 in an hour or of the percentage of total starch gelatinized 

 within the same period. In many instances such figures 

 may be the same, yet there may have been more or less 

 marked differences in the progress of gelatinization dur- 

 ing the early periods of the experiments. Notwithstand- 

 ing such defects, there is in general a remarkable degree 

 of conformity of these curves with taxonomic data. There 

 should be considered with the foregoing the figures pre- 

 sented in Table B 1 which give the numbers of very high, 

 high, moderate, low, and very low reactions ; the sums of 

 reaction-intensities ; and average reaction-intensities of 

 each starch and each parent-hybrid set of starches. 



GENERAL CONCLUSIONS DRAWN FROM RESULTS OF 

 THE HEMOGLOBIN RESEARCHES. 



The results of the crystallographic studies of the 

 hemoglobins indicate: that there is a common structure 

 of the hemoglobin molecule, whatsoever the source of the 

 hemoglobin ; that the crystals of the species of a genus 

 belong to a crystallographic group which represents a 

 generic type : that the crystals of each species of a genus 

 when favorably developed can be distinguished from 

 those of another species of the genus ; that in some spe- 

 cies there may be found one, two, or three forms of hemo- 

 globin, and that this seems to be a generic peculiarity, 



