THE AMARYLLIS AND STAR LILY FAMILY. 1 77 



with pollen from C. zeylanicum. Soon afterwards several mules 

 which had been raised at Mitcham by the Dean himself also 

 bloomed at Spofforth, these having been obtained by crossing 

 C. capense and C. canaliculatum. He also obtained one seed- 

 ling from C. defixum fertilised with pollen of C. speciosum, and 

 another from C. scabrum fertilised by C. canaliculatum, and a 

 handsome white hybrid from C. brevifolium fertilised with 

 pollen taken ffom a large form of C. erubescen's. Dean 

 Herbert; writing in 1842, remarks: "When I first introduced 

 and described a number of species of Crinum which had not 

 been known before in Europe, I was greatly censured by some 

 experienced botanists for asserting that plants which they held 

 to be species at Amaryllis, were in fact variations of the genus 

 Crinum, and it was even declared that Crinum was more nearly 

 allied to Pancratium than to the species in question, I proved 

 the justice of my botanical view of that point by obtaining not 

 merely sterile mules but a fertile offspring between the common 

 Cape Crinum, which was before erroneously called Amaryllis 

 longifolia, and the great Crinum pedunculatum of New Holland. 

 I have now in my garden a further seedling from such a mule, 

 between the Crinum capense and Crinum canaliculatum, which 

 is closely akin to pedunculatum, with ripe seeds upon it. 

 Generally these hybrids become impregnated by the pollen of 

 Crinum capense, of which a great bed stands near them, and the 

 offspring being two-thirds capense, revert nearly to its aspect ; 

 but the plant above mentioned did not revert, but exhibits an 

 improved form of the mule, and is in fact a new fertile species. 

 The freedom with which species of Crinum of the old Linnaean 

 section and most of the section I added thereto interbreed, 

 furnishes decisive proof that the faculty of intermixture is not 

 confined to genera in which species have been rashly formed 

 out of seminal varieties, but is found when the species were 

 even erroneously considered to be of different genera," Nearly 

 all the species of Crinum known to Dean Herbert were by him 

 found to interbreed with tolerable freedom. One unintelligible 

 impediment, however, existed, as he tells us (' Amaryllidacese,' 

 p. 372), for some time. C. capense, which bred freely with 

 every other species, refused to be fertilised by the tropical Cape 

 Coast kinds C. Broussonetianum, C. petiolatum, and C. specia- 

 bile. A seedling was, however, at last obtained between the last 

 named and C, capense. At the page above cited he remarks 

 that, " in general, hybrid plants have been found to be exces- 

 sively florid, but sometimes the contrary has been the case, 

 and there appears to be some impediment to the perfection of 

 their blossoms." At p. 351 he observes that a seed-pod from 



M 



