STANDLEY — ALLIONIACEAE OF MEXICO. 



403 



Leaf blades lance-ovate, twice 

 as long as broad or less; 

 involucres about 6 mm. 



high 10. A. corymbosa. 



Anthocarp not glabrous. 



Leaves linear or narrowly linear-lanceolate. 



Leaves distinctly petioled, comparatively thin. . 13. A. divaricata. 



Leaves sessile, thick and fleshy 14. A. linearis. 



Leaves lanceolate or wider. 



Inflorescence mostly axillary. 



Stems hirsute 15. A. aggregata. 



Stems viscid, not hirsute 16. A. brandegei. 



Inflorescence terminal. 



None of the leaves conspicuously petioled. 

 Plant stout; leaves linear- lanceolate, 55 

 mm. long and 17 mm. wide or less; 

 lobes of the involucre elliptic or 



ovate, obtuse 17. A. pseudaggregata. 



Plant smaller and more slender; leaves 

 linear-lanceolate, 27 mm. long 

 and 5 mm. wide or less; lobes of 

 the involucre lanceolate or ellip- 

 tic, acute 18. A. trichodonta. 



All leaves except the uppermost conspicu- 

 ously petioled. 

 Stems more or less pubescent through- 

 out, densely so above. 



Stems viscid-puberulent below 19. A. coahuilensis. 



Stems subhirsute below 20. A. greggii. 



Stems glabrous at least below. 



Stems woody below (hispid above; 

 ribs of the fruit broad and 



smooth) 21. A. suffrutfcosa. 



Stems herbaceous throughout. 



Involucres glabrate in age, 

 usually 1-flowered, 6 to 

 11 mm. high; ribs of the 

 fruit broad and smooth . . 22. A./oliosa. 

 Involucres densely viscid, 3- 

 flowered, about 6 mm. 

 high; ribs of the fruit 

 narrow and tuberculate. . 23. A. melanotricha. 



1. Allionia coccinea (Torr.) Standley, Contr. Nat. Herb. 12: 339. 1909. 



Oxybaphus coceineus Torr. Bot. Mex. Bound. 169. 1859. 



Mirabilis coccinea Benth. & Hook. Gen. PI. 3: 3. 1880. 



Allionia linearis coccinea Jones, Contr. Western Bot. 10: 51. 1902. 



Type locality: "Hill-sides, Copper Mines," New Mexico. Type collected by 

 Charles Wright, no. 1723. 



Specimens examined; 



Sonora: San Jose" Mountains, 1893, Mearns 1757; Santa Cruz Mountains, Capt.E. 



K. Smith. 



This is so different from other species of the genus as to appear almost anomalous, 



which accounts for the fact that it has long been considered to belong to another 



genus, Mirabilis. It can not well be placed there, however, because it has 2 or 3 



84645°— vol 13, pt 11—11 3 



