Additions to Akboeescent Flora 131 



this tree was first studied. This form is shown in Garden and Forest, 

 10: fig. 39, 5 and 6. The fruit is slightly oblong, compressed, 

 rounded at both ends ; nut compressed, slightly or not angled, rounded 

 or subcordate at apex, pale browTi or mealy ; husk thin, usually split- 

 ting so as. to free the nut. The following forms are sufficiently dis- 

 tinct to merit notice : 



11. PALLIDA PYRiFOTiMis n. c. {Carija p. var. Ashe: Bui. Charleston 

 Mus. 14: 11. 1918). Fruit slightly compressed, ficiform, 2-2.5 cm. 

 long, with a distinct neck ; not obovate, husk very thin not splitting so 

 as to free the nut (Gard. & For. 10: ' fig. 39, 7 and S : 1897). Vir- 

 ginia to Alabama, in the Piedmont section and lower mountains. 



H. PALLIDA AREXicoLA u. c. (C. pallida YSLT Aslic, 1. c). Fruit 

 large 2.5-4 cm. long, oblong but with a stipe, 4-angled by winged 

 husk, nut oblong, more or less angled and pointed. Deep sandy soil 

 in the coastal plain of ISTorth Carolina to Georgia. 



H. PALLIDA APPosiTA u. c. (C. pallida var. Ashe, 1. c). Fruit ob- 

 long to slightly oval, 2-3 cm. long, somewhat compressed, husk 3-4 cm. 

 thick, nut dark brown, angled, resembling that of Ilicoria alba (Gard. 

 & For. 10: fig. 39, 9, 10, 11. 



This tree is extremely close to Ilicoria hucJcleyi (Dur.) n. c. {Carya 

 huMeyi Dur. : Pro. Phil. Acad. Xat. Sci. 12: 547), and fuller investi- 

 gation may show that it is only a variety of //. hackleyi. The chief 

 points of difference seem to be that the fruit of hucMeyi is puborulent 

 and subglobose. The fruit of all forms so far referred to pallida are 

 glabrous. Pubescence of fruit is not considered a constant character 

 in the shagbarks and may not be in this gTOup. The fruits of nearly 

 all species of this genus vary widely in shape. 



HicoRiA viLLosA (Sarg.) Ashe has been regarded as being the 

 same as //. pallida with very slight dift'erencc in fruit, that of villosa 

 as originally figured (Sargent: Silva 7: 1(57) being slightly obovate, 

 the nut also being slightly obovate. The leaves in herbarium sjK'ci- 

 mens show no characters which justify separating villosn from pallida. 

 T]\i' ])late of villosa (Sargent 1. c.) shows tlie pistilhito tiowcr to be 



