318 MODIFIED CIRGUMNUTATIOX. Chap. VH 



the early spring from shoots on some cut-down plants in the 

 greenhouse, slept in a totally different manner from the normal 

 one ; for the three leaflets, instead of twisting on their own axes 

 80 as to present their lateral edges to the zenith, tnmed tipwards 

 and stood vertically with their apices pointing to the zenith. 

 They thus assumed nearly the same position as in the allied 

 genus Trifolium ; and on the same principle that embryological 

 characters reveal the lines of descent in the animal kingdom, so 

 the movements of the small leaves in the above three species of 

 Melilotus, perhaps indicate that this genus is descended from 

 a form which was closely allied to and slept like a Trifolium. 

 Moreover, there is one species, M. messanensis, the leaves of 

 which, on full-grown plants between 2 and 3 feet in height. 

 Bleep like the foregoing small leaves and like those of a Trifolium. 

 We were so much surprised at this latter case that, until the 

 flowers and fruit were examined, we thought that the seeds of 

 some Trifolium had been sown by mistake instead of those of a 

 Melilotus. It appears therefore probable that M. messanensis 

 has either retained or recovered a primordial habit. 



The circumnutation of a leaf of M. officinalis was traced, 

 the stem being left free ; and the apex of the terminal leaflet 

 described three laterally extended ellipses, between 8 a.m. and 

 4 P.M. ; after the latter hour the nocturnal twisting movement 

 commenced. It was afterwards ascertained that the above 

 movement was compounded of the circumnutation of the stem 

 on a small scale, of the main petiole which moved most, and of 

 the sub-petiole of the tej'minal leaflet. The main petiole of a 

 leaf having been secured to a stick, close to the base of the sub- 

 petiole of the terminal leaflet, the latter described two small 

 ellipses between 10.30 a.m., and 2 p.m. At 7.15 p.m., after this 

 same leaflet (as well as another) had twisted themselves into 

 their vertical nocturnal position, they began to rise slowly, and 

 continued to do so until 10.35 p.m., after which hour they were 

 no longer observed. 



As M. messanensis sleeps in an anomaloiTS manner, unlike that 

 of any other species in the genus, the circumnutation of a 

 terminal leaflet, with the stem secured, was traced during two 

 days. On each morning the leaflet fell, until about noon, and 

 then began to rise very slowly; but on the first day the rising 

 movement was interrupted between 1 and 3 p.m. by the formation 

 of a laterally extended ellipse, and on the second day, at the 

 Bame time, by two smaller ellipses. The rising movement thi'n 



