304 



SUMMARIES OF THE HISTOLOGIC CHARACTERS, ETC. 



parent, but in number is closer to the pollen parent ; and 

 the smaller sizes are closer to the seed parent, but the 

 larger sizes closer to the pollen parent. A similar split- 

 ting and shifting is seen in Miltonia bleuana, in which 

 the form is closer to the seed parent; the character of 

 the hilum closer to the seed parent, but eccentricity is 

 closer to the pollen parent ; the character of the lamellae 

 is closer to the seed parent, but certain other features 

 closer to the pollen parent, or as close to one as to the 

 other parent; and the common sizes are closer to the 

 pollen parent. These last two instances are exceptional, 

 probably, merely because of inadequate data. In over 

 half the hybrid is the same as or closer to one parent in 

 only two designations, and in less than half in three 

 designations. In only two are all four designations alike, 

 and in only two are all four designations different, in 

 their parental relationships. 



It is of especial interest to note that in 15 of the 50 

 hybrids (nearly one-third) character and eccentricity 

 of the hilum are separated in their parental relation- 

 ships, character in 1.2 being closer to the seed parent and 

 in 3 being closer to the pollen parent; while eccentric- 

 ity in 12 is closer to the pollen parent and in 3 closer to 

 the seed parent (an exact reversal), a most remarkable 

 peculiarity and one that is very suggestive in connection 

 with the processes concerned in the formation of the 

 starch grain. Another of the several forms of splitting 

 is instanced in Ncrine queen of roses, where the hilum in 

 distinctness is closer to the seed parent, but in fissura- 

 tion, character, and eccentricity closer to the pollen 

 parent ; and it is very much less often fissured but more 

 eccentric than in either parent. The lamellae appear to 

 show less tendency to a splitting of their characters in 

 their parental relationships, but this may be merely 

 apparent and not actual, as will probably be brought out 

 by a sufficiently detailed study. In 5) of the hybrids 

 there occurred an obvious splitting of lamellar properties, 

 this being noted in a separation of character and num- 

 ber; but here, unlike in the case of the hilum, there 

 is not a definite inclination generally of one or the other 

 of these features to one or the other parent. In the split- 

 ting of the hilum into character and eccentricity, charac- 

 ter tends to the seed parent and eccentricity to the pollen 

 parent ; but in the lamellae split, character, and number 

 swing apparently indifferently to one or the other parent. 

 In size, splitting of characters seems to be comparatively 

 uncommon, though here as elsewhere in these studies it 

 is probably not so much an absence of commonness as of 

 careful investigation and analysis. Such splitting as has 

 been recorded under this designation has been manifested 

 chiefly in the ratios of length to breadth of the grains 

 and of the common sizes to other types and different 

 types of grains. 



QUALITATIVE AND QUANTITATIVE REACTIONS OF 

 STARCHES OF HYBRIDS WITH ESPECIAL REF- 

 ERENCE TO REVERSAL OF THESE REACTIONS IN 

 THEIR PARENTAL RELATIONSHIPS. 



(Table E, Parts 1 to 21 and Summary.) 



In the first section, in the tabulations of the starches 

 in regard to histologic and polariscopic properties and to 

 the reactions with iodine and various chemical reagents, 

 data were collected to indicate that the characters em- 



braced under the designations form, hilum, lamellae, and 

 size, respectively, may in each designation collectively be 

 independently heritable ; or that each designation may be 

 split into several independently heritable characters, so 

 that a given hybrid may have a starch that is like 

 that of the seed parent in form, but like that of the 

 other parent in the lamellae; or that it may be like one 

 parent in the general characters of the hilum, but like 

 the other parent in the eccentricity of the hilum, and 

 so on. In the second section, further consideration was 

 given to these peculiarities with reference to histological 

 inheritance. It was shown, moreover, that each reaction 

 is, in its qualitative and quantitative manifestations, 

 heritable independently of each other, so that while with 

 a given reagent there may be sameness or near sameness in 

 the qualitative reaction to the seed parent, with another 

 reagent the relationship may correspond to the pollen 

 parent ; that while a given qualitative reaction may cor- 

 respond to that of the seed parent, the correlative quanti- 

 tative reaction may correspond to that of the pollen 

 parent, etc. ; and that while with one reagent the rela- 

 tionship may be to the seed parent, with another reagent 

 it may be to the pollen parent, and so on. These parental 

 similarities and dissimilarities are of such interest and 

 suggestiveness in connection with both the constitutional 

 peculiarities of different starches and the mechanism 

 of heredity that it seems desirable to tabulate such data 

 more fully and with especial reference to the reversals 

 of the qualitative and quantitative reactions of each agent 

 and reagent in their parental relationships. Of Table E 

 it will be noticed that with only three of the agents and 

 reagents were the reactions of all of the 50 hybrids re- 

 corded; and that in the others the number of hybrids 

 varied from 1 to 32 (in seven less than 10, and in eleven 

 10 or more the restricted numbers being due to the 

 limitations of studies of the qualitative reactions). 



The most conspicuous features of these tables, apart 

 from those already referred to, are : 



(1) The absence in members of a genus of constancy 

 of both qualitative and quantitative reactions as regards 

 sameness of the reactions in their parental bearings ; 

 (2) the tendency to the appearance of a definite ratio 

 in the qualitative reactions in their inclinations to the 

 seed and pollen parent; (3) the tendency to an absence 

 of such a ratio in the quantitative reactions in their in- 

 clinations to the seed and pollen parent; (4) the large 

 percentage of instances of reversal of the parental rela- 

 tionships of qualitative and quantitative reactions with 

 given agents and reagents. 



It will be noted that in the reactions with each rea- 

 gent there does not exist generic constancy or uniformity 

 of either qualitative or quantitative reactions in their 

 parental closeness. For instance, while in the chloral 

 hydrate qualitative reactions of Brwwdonna, Hippeas- 

 trum, Hcemanthus, and Begonia all of the hybrids be- 

 longing to each genus incline to the seed parent, in all 

 other genera represented in which there are two or more 

 members some of the hybrids of each genus incline to one 

 parent and others to the other parent. Thus, in Crininn 

 one hybrid inclines to the seed parent and two to the 

 pollen parent ; in Nerine four incline to the seed parent 

 and one to the pollen parent; in Narcissus eleven incline 

 to the seed parent and two to the pollen parent; in 

 Lilium three incline to the seed parent and two to the 



