70 PRINCIPLES OF PLANT-TERATOLOGY. 



In many double roses a great increase in the corolla-, 

 whorls occurs, as also in double tulips, Goebel cites 

 many cases, as in the blue lobelia (Lobelia Erinus), in 

 which the earliest-formed whorls of petals were regu- 

 larly 5-merous, but the later-formed ones irregular. 



Lagerberg describes a flower of Viola, mirabilis in 

 which there was an extra inner whorl of three petals 

 alternating with the normal ones, and bearing spurs : 

 a case of partial pelory. Here we see the ternary 

 character appearing in the corolla, which is a normal 

 feature in the gynoeceum. 



Masters mentions a Gyclamen flower which had 

 several extra whorls of petals; as there .were no 

 stamens, one of these whorls must probably be due to 

 their petalody, the rest to dedoublement. 



Interesting examples of the formation of extra whorls 

 are afforded by a species of thorn-apple [Datura fas- 

 tuosa). The simplest case is that in which there is an 

 extra corolla, the rest of the flower being normal. In 

 other cases, as cited by Goebel, the normal calyx and 

 corolla were 5-merous ; then followed a second corolla 

 of about ten petals, being succeeded in its turn by a 

 third and fourth corolla or even a greater number; the 

 androecium and gynoeceum were imperfectly formed. 

 Goebel declares that this is not a case of dedoublement, 

 as the extra whorls arise independently and alternate 

 regularly ; this statement, however, is made merely 

 from the point of view of developmental data to which 

 he attaches, as ever, an exaggerated importance. 

 Knowing as we do that in other plants extra whorls 

 of petals do arise by fission of the normal petals, it is 

 highly probable that in Datura the extra whorls have 

 arisen in the past in the same way, but now arise 

 congenitally as quite distinct whorls, and, arising 

 separately in this way, it has become possible for the 

 members of the whorls to be rearranged so as to 

 conform with the law of alternation. An intermediate 

 stage, taking us one step nearer the ontogenetic origin 

 of the phenomenon, is figured by Masters for the same 



