(7) FREDERIC NEWTON WILLIAMS. UENRE RH0DALS1NE GAY. 7 



tain other characters which in the aggregate may serve to group the 

 known forais under two species. After the fnll description which has 

 been given of this species, it is only necessary to indicate the points of 

 différence between the next species and this : so that when spécimens 

 are laid side by side they may be readily distinguishable. 



GEOGRAPHICAL DISTRIBUTION 



Spaix. Prov. of Valentia, San Felipe («Arenaria extensa », Du four [1820]). 

 Prov. of Murcia, Sierra de las Cabras, near Hellin (in Herb. Kew. n. 84). 

 Italy. Sicily, Mt. Etna (Strobl 1885). Sardinia (Moris, herb.). 

 Algeria. Oran. 1866 (Munby, in Herb. Kew.). 

 Morocoo. Taurirt, al 3000 mètres, 1888 (./. Thomson, in Herb. Kew.). 



2. Rhodalsine geniculata, Williams. 



Haec altera species a prsecedenti notulis sequentibus difiert. — Spe- 

 cimina plerumqne robustiora, indumento valde variantia. Petala alba. 

 Stamina inaequilonga, longiora 2 / 3_3 A longit. petalormn attingentia, bre- 

 viora (petalis opposita) usque ad medium petalorum vix attingentia. 

 Capsula fructifera calycem paullum excedens. Semina tuberculata (oculo 

 nudo), plerumque 8-11. 



a. Poiretiana, Williams. 



Planta inferne glabra, eglandulosa, superne villoso-puberula glandu- 

 losa. Folia lineari-oblonga. Sepala 5-7-nervia. 



Stjn. Arenaria geniculata, Pair. Voy. Barb. II, p. 166 (1789); et Eneycl. 

 Meth. VI, p. 365 (1804); Jacks. Ind. Kew. I, p. 179 (1893), 



Alsine proeumbens, Battand, et Trab. Fl. de l'Algérie, I, p. 154 (1888). 



GEOGRAPHICAL DISTRIBUTION 



Algeria. (Poiret, the original spécimens). Oran, 1839 (Battandier). (Bové iu 

 Herb. Kew.). 



Gay has the following note attached to Bové's spécimen, which is 

 labelled « Arenaria herniarisefolia, 1839, » and from the character of 

 its indumentum is readily referable to the type : — « l'échantillon de 

 Boissier est trop jeune ; dans celui de Bové, qui est un peu plus avancé, 

 je n'ai pu découvrir qu'une seule capsule en état de déhiscence, et cette 

 capsule était à 5 valves comme Desfontaines et Poiret le disent; et non à 



