OJt' NORTH AMERICA. 49 



633 revised it and separated by the unilocular 

 berry, reducing to SARCOCIDIA the Genera Phy- 

 tolaca, Raxamaris, Schefferia &c, and to Ri- 

 VINIDIA the pretended Chenopodea with a ber- 

 ry. The EMPETRIDES are more akin to those 

 than to Euphorbides and Celastrides, and are 

 quite unlike to Conifera. The genera belong- 

 ing thereto are Grubbia, Coilosperma, Batis, 

 Ceratiola &c besides the types Empetrum and 

 Colema- . . . but the Genera Skimmia, Nandi- 

 na, Melicytus &c are very near also. The 

 whole requires a careful revision. 



594. COLEMA Don 1826. Emfcucum Raf. 

 1886 fl.'tel. Dioical, calix 5-6parted, stamens 

 3 to 4, style 3-4parted, berry 8-41ocular, 3-4 

 sperm. Evergreen shrubs, habit of heaths, 

 Jlowers terminal and axillary glomerate 

 This had been based on the Empetrum album 

 of L. but our sp. may be of quite a different 

 Genus, the real Colema having the ternary part 

 constant, no style but 3 stigma, a berry nearly 

 monolocular when ripe. Therefore it must 

 form the subgenus Euleuca, and our sp. with 

 heterogonal parts and stamens, a style, and pro- 

 bably a 41ocular berry must form a subgenus 

 ENDAMMIA Raf. if not a Genus, meaning in 

 sands. 



595. COLEMA ARENARIA Raf. or Endammia 

 ericoides Raf. Empetrum conradi, Torrey 1835. 

 Cespitose procumbent smooth, leaves subverti- 

 cillate and alternate, narrow linear acute glan- 

 dular, margin re volute, flowers glomerate and 

 capitate in the sandy tracts among the Pine 

 woods of New Jersey, first noticed by Kin in 

 1800, who gave it to me as an American heath 

 in 1802, found by me in 1804 near Pemberton, 

 long before Conrad, and twice again in 1833 at 



7 



