8 



INTRODUCTION. 



rum (parent) 3/x and the average 1.5^; and in R. cul- 

 verwellii (hybrid) 5/n and the average 3/t. In Menziesia 

 empetriformis var. the largest starch grains are 6/x, and 

 in all cases they are larger than in the other parent 

 Rhododendron chamcecistus; while in the hybrid Bryan- 

 thus erectus the grains are 4/t across at their largest, 

 though most are from 2 to 3/*, the size being intermediate 

 but falling rather toward the latter parent. Macfarlane 

 states : 



" Hedychium gardnerianum, the one parent of H. 

 sadlerianum, forms strong rhizomes, whose storing cells 

 are large, but scantily filled with starch in all that I 

 have examined. Each starch grain is a small, flat, trian- 

 gular plate, measuring 10 to 12/u. from hilum to base, 

 and the lamination is not very distinct. H. coronarium, 

 the other parent, forms smaller and fewer rhizomes, 

 and the starch-storing cells are from half to three-fourths 

 the size of the last, but these are densely filled, particu- 

 larly in the central parenchyma, with large starch gran- 

 ules. Each is ovate, or in some cases is tapered rather 

 finely to a point at the hilum. They are from 32 to 

 60/i long, measuring as before, and the lamination is 

 very marked. The cells of the hybrid are on the average 

 between those of the parents; but if one may judge by 

 opacity of cells the amount of stored starch approaches 

 more closely to that of the latter parent. The grains 

 may best be described if we suppose a rather reduced one 

 of the first parent to be set on the reduced basal half of 

 one of the latter. The lamination also is more pro- 

 nounced than in the first, less so than in the second. 



" A second cross was effected by Mr. Lindsay with 

 //. coronarium, and examination of the rhizome starches 

 proves that the second hybrid approaches very closely to 

 the species parent. But the grains of H. lindsayi illus- 

 trate microscopically a phenomenon which has been re- 

 peatedly referred to, viz, the greater variability and 

 instability of a second over a first hybrid ; for many of 

 the grains (in some specimens the majority) have fantas- 

 tic shapes, appearing as if undergoing rapid disintegra- 

 tion by leucoplasts, or perhaps more truly as if the latter 

 were incapable of building up the shells of starch in a 

 regular and uniform manner. 



" A set of crosses has been effected between H. elatum 

 and //. coronarium. The grains of the first are like those 

 of PI. gardnerianum, except that they are larger (18 to 

 24ft), and that the lamination is coarse. The grains of 

 the hybrid are larger than those of H. sadlerianum, and 

 exhibit even more evident lamella?. They measure on the 

 average, 40/x, but vary from 30 to 50/i. Not infrequently 

 all the above hybrids have (mixed up with grains more 

 typically intermediate) some grains which can scarcely, 

 if at all, be distinguished from the small ones peculiar 

 to one parent, while very rarely I have observed grains 

 so large and rounded as to pass for those of //. coro- 

 narium. Now, when describing the epidermal leuco- 

 plasts of Dianthus grievei it was stated that, though the 

 average was nearly 3ft, some measured 2.5/i or slightly 

 less, others as much as 3.5ft. The occurrence of these, 

 and similar minute differences in protoplasmic masses, 

 or in formed materials like starch grains which are due 

 to manufacture by these masses, induced me to prepare 

 a set of micro-photographs, and to project lantern trans- 



parencies of these on a 7-foot screen. Thus it was pos- 

 sible to study their dimensions more exactly than under 

 the microscope. It was then found that while the shape, 

 appearance, and size of most starch grains of Hedychium, 

 of Dianthus leucoplasts, and of Geum and Masdavallia 

 chromoplasts were intermediate, examples might be got 

 which reverted powerfully to one parent, and, so far as 

 they have yet been studied, the reversion was most fre- 

 quently towards the parent with the more minute cell- 

 contents." 



The results of the studies of starches are therefore 

 in entire accord with Macfarlane's conclusions pertaining 

 to the tissues in showing intermediateness of the hybrid, 

 with a tendency at times to a leaning to one parent. 



Investigations of the starches of varieties and of 

 parents and hybrids of varieties of round and wrinkled 

 peas have been made by Gregory (The New Phytologist, 

 1903, n, 226), Weldon (Biometrica, 1902, i, 246), and 

 Darbishire (Proc. Roy. Soc., B., 1908, LXXX, 122; Breed- 

 ing and the Mendelian Discovery, 1912, 124). 



Gregory (The New Phytologist, 1903, n, 226) found 

 that the starches of round and wrinkled peas occur in 

 two very different types. In the round seeds the periph- 

 eral cell-layers of the cotyledons contained a few oval 

 starch-grains which did not exceed 0.06 mm. in the great- 

 est diameter. In the third layer the grains reached 0.2 

 mm. in length, while the more deeply situated cells were 

 crowded with oval grains measuring as much as 0.34 mm. 

 in the greatest dimension. The grains were regular in 

 shape, with a definite center surrounded by well-marked 

 lines of stratification. In the wrinkled peas the grains 

 of the peripheral layers were of about the same size as 

 those of the round peas, but were of a different type, 

 occurring in irregular spheres with several centers, thus 

 forming a compound grain which has a strong tendency 

 to break up into smaller parts. In the cells which lie 

 deeply these compound grains never attain a greater 

 length than 0.1 mm. in the greatest dimension. Table 1 

 gives a list of the seeds examined. 



TABLE 1. 



Gregory notes that seeds of intermediate and dubious 

 shapes were not uncommon in certain of the races. The 



