The Scottish Naturalist. 



species of Valeriana above-mentioned ; the affected margins were 

 furnished with pectinately-arranged sharp teeth. In Valerianella 

 Szovitsiana the rolling of the margin was less striking ; the affected 

 portion of the leaf developed into a prolonged lobe, and was pro- 

 vided with small notched teeth. In both species the abnormalities 

 just described occurred usually on only one pair of leaves. The 

 next pair showed them in less degree, if at all, and the succeeding 

 leaves and the flowers remained quite normal. Had the infection 

 been made at the time when the flowers would have been fully 

 exposed to it, I should certainly have obtained double flowers in- 

 stead of abnormally-formed leaves. I have exposed V. olitoria to 

 copious infection at successive intervals ; and I obtained 

 abnormally-toothed leaves and double flowers. 



The various Valerianacece, at least the species of Valeriana, 

 which I have most fully studied, exhibit various specific peculiari- 

 ties. The species with thin leaves are more easily infected than 

 those with thicker leaves. Patrinia rupestris, the leaves of which 

 are rough to the touch, remained intact. Valeriana saxatilis and 

 celtica, in which infection often failed, developed, as already men- 

 tioned, only abnormal leaf-forms, but not double flowers, only in 

 the former there occurred metaschematic flowers ; in V. elongata 

 in each experiment I always obtained abnormally formed leaves, 

 but no double flowers ; the few flowers that showed themselves 

 developed ill ; apparently I transferred too many individuals of 

 Phytoptus in these experiments. In Valeriana Phu, which was 

 often infected, I never obtained strikingly abnormal leaf-forms, but 

 once indeed double and proliferous flowers. In this case individual 

 bracts were lobed and also petaloid. Without doubt one must 

 obtain double flowers in V. alliarioefoUa and saliunca if the 

 infection is made at the proper time. Centranthus ruber yields to 

 infection only with more difficulty than the two annual species, (7« 

 Calcitrapa and macrosiphon. The abnormalities that occurred on 

 the infected G. ruber were rather varied. Too strongly infected 

 examples were crippled ; some showed rolled leaf-margins, but 

 without the characteristic subsequent development of pectinate 

 teeth ; I obtained metaschematic flowers, flowers with several 

 spurs, and double flowers. The Phytoptus does not remain on 

 either V. Phu or C. ruber. Infection of V. Phu in autumn was with- 

 out result. 



The Cruciferse are not very well suited for the nutrition of our 



