crimson, brick-dust red, orange and bicolor (crimson and 

 white) ; the latter plant seemed to me to be more luxuriant 

 where the soil had been broken by the liberated Africans for 

 the purpose of cultivating the Manioc ; the other varieties 

 become scarce where the soil had been disturbed, but in 

 every instance, except the first plant, I found it growing in 

 what I consider a strong loam, impregnated (from its colour 

 and aptitude to stain linen) with a large portion of carbonate 

 of iron, 



" The plant Mr. Knight has flowered, being scarlet, 

 makes five distinct colours ; it was brouo^ht bv me from 

 Sierra Leone in May, 1840, but I had previously sent to the 

 Duke of Bedford all the varieties I have named, and I had 

 hoped that they would have flowered long before Mr. 

 Knight's, as they had had twelve months advantage, having 

 been sent to His Grace in March, 1839, and again in July of 

 the same year. 



*' When the root is shaded from the sun by the under- 

 wood, this C. splendens attains the height of 10 to 12 feet, 

 but if it is exposed to the sun it seldom grows more than 

 three feet." 



The varieties here spoken of are probably distinct spe- 

 cies, of which several from the same part of Africa are known 

 to Botanists. 



Judging from the general character of the Sierra Leone 

 climate, it may be expected that this species will require con- 

 siderable bottom heat, a hot moist atmosphere while growing, 

 and at least four months rest in a drier atmosphere after its 

 growth is completed. 



Upon looking over the characters assigned to the genus 

 Clerodendron, I have not found any mention of the singular 

 structure of its ovary, which in this, and some other species, 

 is as is represented at fig. 1. of the accompanying plate. 

 That is to say, it is one-celled, with two bifid parietal pla- 

 centae, which project into the cavity, having two ovules on each 

 edge, and touching in the middle. At the same time there is 

 an imperfect partition on each side, slightly projecting into 

 the cavity opposite the commissure of these placenta?. The 

 learned reader need not be reminded of the close approach, 

 thus exhibited, to the structure of Martvnia. 



