MINUTE STRUCTURE OF PLANT HYBRIDS. 



225 



examples looked at, since they are 8 to 9 m in diameter, and have each a thick wall 

 with shrivelled-looking contents. I regard the pollen, therefore, as being entirely impotent, 

 and pollination experiments which have been conducted repeatedly, verify this by their 

 negative results. 



Nectary. — There is an extremely neat nectar- secreting arrangement worthy of study. 

 The pollen parent (3) — a less specialised form in every way to my mind — shows this in 

 its simplest state as a small saucer-shaped receptacular girdle (Plate IV. fig. 14, c) between 

 the stamens and pistil. The entire depth of the gland tissue is about 180 m. In the seed 

 parent a special contrivance has been devised to protect and economise the nectar, for in it 

 the bases of the petals have fused with the receptacle, and this has got deeply excavated 

 round the ovary to form a nectar ditch the walls of which are lined by gland tissue 

 supplied by a Special set of subjacent bundles. The depth of gland tissue is 400 to 

 420 m (fig. 14, a). In the hybrid there is an exactly intermediate state of things, in the 

 fusion of the petals, shape and size of the gland, and depth of its tissue (fig. 14, b.) 



1. Geum rivale. 



2. Geum inter- 



medium. 



3. Geum urbanum. 



(c) Geum intermedium, x . 



This hjdmd has been regarded by eminent systematists as a true species. Probable 

 reasons for this are its extreme frequency of occurrence, the large number of good pollen 

 grains, and the usual production of abundant seed. Many suspected its hybrid nature, 

 which has been experimentally verified by Gartner and Bell Salter. To speak only of 

 the Edinburgh neighbourhood, it is frequent where both parents are found, and is usually 

 mixed with these. Thus, in Colinton Glen, Mr Rutherford Hill has gathered quantities ; 

 in Carriber Glen, near Manuel, it is very abundant, and has been carefully watched by me 

 on repeated visits ; in the valley of the Esk it also appears frequently ; but by far the 

 finest locality was pointed out to me by Mr Hill. The woods along the shore in the Dal- 

 meny policies beyond Cramond have large areas covered by it alone, or shared also by one or 

 both parents. It is easily distinguished by naked eye from the size of the stipules 

 developed by the upper cauline leaves ; by the size, position, and colour of the flowers, 

 while — as Godron * and others have pointed out — the hybrid is commonly more luxuriant 

 than either parent. 



Diverse in naked-eye appearance though the parents are. I have found more difficulty 

 in seizing on minute histological differences than in any other set. Marked changes in 

 some parts, however, have been effected, which appear in blended manner in the hybrid.t 



Root. — The mature root in all three exhibits a broad annual cork formation, and a 

 very rare feature in plants is the formation in all of intercellular spaces between the cork 

 cells. In 1 \ the growth of each annual cork ring proceeds so that the oldest brown layer 



* Mem. Acad. Stanisl., 1865, p. 347. 



t Throughout the description I have assumed Geum rivale to be the seed parent, and G. urbanum the pollen parent; 

 To select two plants, however, to protect these, and to raise offspring from reverse crosses for comparison, is most desir- 

 able, and our knowledge cannot be exact till this is done. , 



X The growth formation of these requires further study. 



