GIBRALTAR PLANTS 11 



specimen in his herbarium has them subcordate at the base of the 

 limb. Most authors, however, have regarded them as eUiptical 

 and more or less narrowed below, and from this, together with 

 the synonymy cited, much confusion has arisen. Boissier (Voy. 

 Bot. pp. 12-13) recognized that in this species and its allies or 

 varieties the lateral petals vary indefinitely in form, and his 

 D. peregrin uvi covered species and varieties with them either 

 orbicular and subcordate, or elliptical and more or less narrowed 

 below. DeCandolle (Fl. Fr. vi. p. 641) appears to have been the 

 first to describe a species, D. junceum DC, having elliptical 

 lateral petals, his D. cardiopetalum and D. gracile having them 

 suborbicular. In addition to the form of the petals, the aggregate 

 species varies indefinitely in leaf-cutting and habit, the racemes 

 being sometimes dense and compact, and at others elongate and 

 lax, while the leaves are firm or flaccid, close-set or distant, so 

 that a subdivision into fixed species has not proved satisfactory. 



Perez Lara (Fl. Gad. p. 89) observed that D. longipcs Moris 

 (Z). peregrinum var. longipes Boiss.) always has suborbicular 

 lateral petals, and proposed the following arrangement : — 



D. peregrinum Linn, a genuinum (D. peregrinmn Boiss.). 

 Lateral petals elliptical, attenuate at base, the racemes either 

 short or elongate [D. junceum DC). 



(i cardiopetalum (Z). cardiopetalum DC). Lateral petals sub- 

 orbicular, truncate at base. 



Sub var. longipes {D, longipes Moris). Lateral petals orbicular, 

 subcordate. 



Subvar. gracile [D. gracile DC). Lateral petals ovate, cordate. 

 This arrangement has the defect of restricting typical D. ijcre- 

 grinum Linn, to a form with elliptical narrowed lateral petals, 

 which its author did not intend, also of making no substantial 

 difference between subvarieties longipes and gracile, which seem 

 to me identical. 



A better arrangement, which I am following in my Flora of 

 Gibraltar, is that of Boissier (I. c), though the definition of his 

 aggregate species and in part the synonymy are my own, viz. : — 

 D. peregrinum Linn. Lateral petals suborbicular, truncate or 

 subcordate at base, or elliptical and more or less narrowed below. 

 Var. confertum Boiss. (Z). cardiopetakmi DC, D. halteratum 

 Sibth. & Sm.). Racemes dense. 



Var. longipes Boiss. (D. longipes Moris, D. junceum DC, D. 

 gracile DC). Racemes elongate, lax. 



In the neighbourhood of Gibraltar I have only seen the 

 form of var. longip)es with truncate or subcordate lateral petals, 

 which is very common in all sandy places, but the following are 

 recorded : — 



D. gracile DC Uncultivated sandy places and cornfields at 

 Algeciras, Beverchon. 



D. peregrinum Linn. var. confertum Boiss. Near Algeciras, 

 Schott. 



Var. longipes Boiss. Catalan Bay and on the Neutral Ground, 

 Boiss., Kelaart, Dautes, and on the slopes of San Roque, Boiss. 



