22li PUCCINIA 



as regards Pol. viviparum Kiebahn expresses doubts, but Semadeni was able 

 successfully to infect that species with uredospores from Pol. Bistorta. The 

 form mi Pol. viviparum is here kept provisionally distinct, since the 

 secidium on Angelica has not been found in Britain. 



Of the second group, P. mammillata, there are two biological races — 

 (1) P. Mei-mammillata Semadeni, on Meum, and (2) /'. Angelicae-mam- 

 millata EOebahn, on Angelica. Neither of these has beep found in 

 Britain. All the four of these are closely allied ; in Sydows' Monographia 

 it is suggested that possibly in all these cases the secidium on the 

 Umbellifer is merely facultative and the Puccinia can maintain itself 

 without that aid. 



96. Puccinia Polygoni-vivipari Karst. 



Puccinia Bistortae DC. ; Cooke, Grevillea, ii. 161 ; Micr. Fung. p. 204 

 p.p. Plowr. Ured. p. L92. Sacc. Syll. vii. 0.38. Sydow, Monogr. 

 i. 571 p.p. 



P. Polygoni-vivipari Karst. Enum. Fung. Lapp. p. 2-2]. Fischer, 

 Ured. Schweiz, p. 100, f. 76 (?). 



U redospores. Sori hypophyilous, scattered, small, roundish, 

 cinnamon-brown, girt by the erect epidermis ; spores roundish 

 to ellipsoid, finely echinulate, pale-brown, 20 — 23 x 16 — 17 /a. 



Teleutospores. Sori similar, but blackish-brown ; spores 

 elliptical to obovate-oblong, rounded and not thickened above, 



Fig. 174. P. Polygoni-vivipari. a, two teleutospores from liar Lodge; 

 b, uredo- and teleutospore from a specimen issued by Fuckel, Symb. Myc. 

 57 ; both on Polygonum viviparum. 



hardly or not at all constricted, rounded or slightly tapering 

 below, smooth, brown, 20 — 30 x 15 — 20/z; epispore very thin and 

 translucent ; no papilla on the germ-pores ; pedicels deciduous, 



short. 



On Polygonum viviparum. Very rare. Xear Mar Lodge, 

 August, 1822 (Dr Greville). Braemar, August, 1882 (Prof. 

 Trail). (Fig. 174.) 



This species is closely allied to P. Conopodii- Bistortae (q. v.), but in 

 the absence of all biological information is best kept distinct. I have 



