228 DR J. M. MACFARLANE ON THE 



/. Geum nvaie. surface. In 3 the elongated basal cells are 26 a. lone;, but they become shorter upwards, 



;.'. Geum inter- « ~ °' » x 



medium. ^^}} they are 15 ju, by 12 a. Their walls are sharply zigzag with fine infol dings. In the 



X Geumurbanum. J r J r r^oo^ ^ o 



hybrid all of these features are intermediate, while the cuticular strise of one parent just 

 described are more finely reproduced in the hybrid. 



Stamens. — The outer anther coat of 1 consists of cells of very irregular outline. 

 The walls form knob-like infoldings, while the cuticle is raised into prominent wavy ridges 

 so strongly developed as almost to obscure the cell outlines beneath. In 3 the cells are 

 roughly quadrangular, and their walls have only minute infoldings at the angles. By 

 careful focussing one can detect a very delicate ridging of the cuticle. In all of these 

 points the hybrid is as exactly intermediate as one can detect. 



The pollen grains of 1 are 26 to 28 a, across ; those of 3 are 18 to 20 a ; and those of the 

 hybrid 26 to 28 a. Illustrations of these are given in Plate V. figs. 6 a, b, c. A noteworthy 

 feature is the very large percentage of good, sound-looking grains produced by the hybrid. 

 The percentage, in blossoms examined from various localities, has fallen as low as 45 to 50 

 percent., but in the majority of cases it averages from 85 to 95 per cent., which is as high 

 as one usually finds in either parent. 



Pistil. — From the opening of the flower onwards the pistil is an interesting study, but 

 I shall only draw attention to the figures on Plate V. figs. 2, 3, 4, which illustrate well their 

 external characters. An explanation of the peculiar relationship of the style to the style- 

 arm must be possible, but it has not occurred to me yet, nor can I get any account of it. 

 That the style-arm must be functional at the time of pollination, or between then and full 

 ripening of the fruit, appears almost certain since each falls off previous to fruit dis- 

 semination. The formation of a projecting knob on the style-arm of 1, of a similar knob 

 on the style-tip of 3, and of one reduced in size on both style-tip and style-arm of the 

 hybrid is interesting, but their use is still a puzzle. 



The embryos differ in their shape as do the outlines of their fruit walls. In the 

 cells of the three embryos proteids and oil are the reserve materials. In 1 the 

 amount of oil is relatively small ; in 3 it is extremely abundant, and exudes from the 

 cells as large refractive globules. In # the amount is considerably less than in 3, but more 

 abundant than in 1. 



The number of achenes that mature their seeds is very great, and this agrees with the 

 high quality of the pollen. The Rev. C. Wolley Dod, however, has stated that with him 

 G. intermedium rarely seeds, though other hybrids of Geum do. But the bulk of evidence 

 goes to show that the hybrid has a favourable combination of circumstances for its perpetua- 

 tion. Growing in shady places frequented by insects, it attains there a high stature, com- 

 bining the strong rank growth of 1 with the elongated branching inflorescence of 3. It 

 begins to blossom at a time exactly intermediate between the parents, and therefore has 

 a great advantage over 3, which has often to struggle with tall-growing summer weeds 

 that rush up later than does the hybrid. The half-erect position of the flowers, with open 

 spreading sepals and large yellowish-red petals, give it a decided advantage over either 

 parent. It is not at all surprising, therefore, that it is so abundant, and, as at Cramond, 



