158 THE JOURNAL OF BOTANY 



Ilysanthes micrantha, sp. nov. Annua, parva, glabra, caule 

 gracili simplici erecto sursum sparsissime ramose distanter folioso, 

 foliis sessilibus anguste linearibus obtusis integris, floribus paucis 

 axillaribus vel pseudoterminalibus, pedicellis sat longis filiformibus 

 erectis sub fructu patenti-nutantibus, calycis abbreviati campanu- 

 lati adusque medium divisi lobis lanceolatis acuminatis, corollae 

 parvas tubo calycem paullulum excedente lato superne brevissime 

 amplificato labio postico anticum breviter excedente late oblongo 

 apice rotundato ipso emarginato labii antici lobis inter se sub- 

 sequalibus rotundatis, staminum posticorum loculo altero minimo 

 casso loculis ambobus basi barbatis, staminodiis magnis superne 

 clavatis necnon longe ciliatis, capsula oblonga acuminata calycem 

 multo excedente. 



Hab. Ehodesia, Victoria ; C. F. H. Monro, 1031. 



Planta summuni 6 cm. alt. Folia + 5 mm. long., circa •5 mm. 

 lat. Pedicelli solemniter 1-1'5 cm. long., sub flore tenuissimi. 

 Calyx totus 2 mm. long. ; tubus 1 mm. long., ore totidem lat. ; 

 lobi 1 mm. long. Corollge tubus 2-5 mm. long., faucibus segre 

 2 mm. diam. ; labium posticum 4 x 2'5 mm. ; labii antici lobus 

 intermedins 3x4 mm. Stamina corollge faucibus infra os fere 

 1 mm. inserta ; filamenta complanata, vix 2'5 mm. long. ; antheras 

 fertiles ovatte, -75 mm. long. Staminodia leviter cohserentia, in- 

 curva, fere 3 mm. long., horum clava 1 mm. long. Stylus 3 mm. 

 long. ; stigma '5x1 mm. Capsula 7-9 mm. long. 



The plant is remarkable for the large club-shaped, coherent, 

 ciliated staminodes and the rudimentary second cell of its stamens. 



(To be continued.) 



NOTES ON EPILOBIUM HYBKIDS. 

 By E. H. Compton. 



The study of naturally occurring hybrids is at^ the present 

 day beset with difficulties, the chief of which is the lack of ex- 

 perimental data. The process of identification of a supposed 

 hybrid is in nearly all cases based on the assumption that the 

 characters exhibited are intermediate between those shown by the 

 parents. _ The species selected as the supposed parents are usually 

 those which differ from the plant under consideration by about 

 an equal amount on either side. As a rule, no experimental 

 evidence is forthcoming, and the chief criterion of hybrid-nature 

 is intermediateness between a pair of species selected to account 

 for the facts. This assumption is carried still further in the case 

 of "tertiary hybrids," which are supposed, on grounds of inter- 

 mediateness, to have arisen from the union of a hybrid with a 

 pure species. In fact, it appears that the words " hybrid " and 

 "intermediate" are sometimes used indiscriminately — a deplor- 

 able, though almost logical, result of the method of identification 

 adopted in the past. When we read in a Flora, " A large number 



