146 ASCOMYCETES. 



but it is completed by the weight of the broom itself, the 

 excessive development of sleeping buds, and the premature 

 death of twigs. Smith also investigated the anatomical changes 

 occuring in witches' brooms due to E.<'ii*-n/<. From his 

 resumd we select the following: "In a witches' broom the 

 increased thickness of the twipjs and branches is due to a 



<_? 



proportionally greater increase in the bark than in the wood, 

 the hypoderm, especially, having its cells more numerous and 

 larger, while their normal arrangement in longitudinal rows 

 is lost. The cork-cells are enlarged and retain their plasma- 

 content longer. The phelloderm is better developed. In the 

 sclerenchynia-ring, the primary bundles of bast-fibres are smaller 

 and further apart from each other, or they may be quite 

 absent ; the bast-fibres are shorter and have thinner walls ; 

 sclerenchymatous cells are more numerous, larger, and have 

 thinner walls. The phloem is increased chiefly through enlarge- 

 ment and increase in number of its medullary rays; phloem 

 crystal-deposits tend to be multiplied. In the wood, the parts 

 most enlarged are the pith and medullary rays ; tracheae are 

 more numerous, but their component elements are shorter ; the 

 wood-fibres have thinner walls, wider lumina, and are often 

 chambered ; the normal course of the long elements is much 

 disturbed by the greatly enlarged medullary rays. 



Sadebeck has recently divided the parasitic E.coasceac into 

 these genera : (.) Magnusiella, with asci isolated on the ends 

 of mycelial threads which lie between the epidermal cells ; in 

 the other genera the asci arise from a subcuticular hymenium : 

 (b) Tnjilirina, without a perennating mycelium ; (c) E,>:<><ixciis, 

 with ;i perennating mycelium ; (//) Taphrinopsis may be taken 

 as another genus. Ascow //<-fx he does not reckon with the 

 Exoasceae. 



Brefeld divides the family into E.t-oascus, with eight spores in the ascus, 

 and T<i./>ln-!.n<i, with fonr-spored asci. Sadebeck shows, however, that 

 eight is the normal number of spores in all the species, and that variation 

 therefrom is frequent, four or more spores or numerous conidia being 

 funned. 



Schroeter separates the genus MiHjumiella, as Sadebeck has done, then 

 divides the remainder into E.roas<-nx with eight-spored asci at time of 

 maturity, while those with many-spored asci are placed under Tavhria 

 (the older name given to TapJmna) 



