NOTES AND OBSERVATIONS. 155 



very sparingly rusty stellate, green, marked with minute 



,pale dots. 



The five petals are oblong, half an inch wide and as 

 much as 7 in. long, whitish green with a shade of chocolate, 

 and microscopically stellate pubescent, the hairs of the 

 outer side being browner than those within, which give a 

 satiny appearance to this side of the petals. On opening, 

 the petals become coiled backwards and expose the pure 

 white stamens as the showy part of the flower* 



The androecium is about 6 inches long, with a thick cor- 

 rugated tube 2 in. long, parted half an inch above the base 

 into ten sets of about eight filaments each, which fork 

 near the middle — each arm ending in a single anther 

 •cell. The anthers are oblong, versatile but nearly erect, 

 and deep yellow, like the pollen. The ovary is conical, 

 tomentose, 5-celled, and tapers into the slender white style, 

 w T hich about equals the stamens and is divided at apex into 

 five short linear yellow stigmas. 



The ovoid fruit becomes about 2x3 inches, and breaks 

 into five valves, each carrying a septum on its middle, and 

 leaving as many rows of about five seeds each attached to 

 an axile 5-winged columella. This, like the inside of the 

 valves, is densely white tomentose, while the seeds are 



glabrous, brown, with a series of somewhat anastomosing 

 elevated white bands radiating from the hilum and meeting 

 over the apex of the seed. 



Martius, in volume i. of the Flora Brasiliensis, gives 

 descriptions of the species of Pachira or Carolinea recog- 

 nized as occurring in Brazil. In his General History of 

 the Dichlamydeous Plants, Don characterizes all of the 

 species recognized in 1831; and Spach, in Hist. Nat. des 

 Vegetaux, iii. 423, also treats of the genus in 1834, since 

 which time no general account appears to have been written, 

 though a number of plates have been published. The 

 descriptions as a rule are incomplete in one or other impor- 

 tant respect, so that it is impossible to name a specimen by 

 them with certainty ; but the number, apex, and degree of 



