592 R. J. HARVEY-GIBSON AND MINNIE BRADLEY ON 



tunned. A large number of cords are then given off' to supply the sepals and 

 petals ; the ring then re-forms, and almost immediately gives off numerous bundles 

 to supply the stamens. Once more the ring closes, and then finally splits into as 

 many vascular cords as there are placental plates. Between these cords there is 

 at first no vascular tissue, but later on independent groups of tracheids are differen- 

 tiated, branching in all directions, but especially towards the ovarial cavity. 

 Further, the placental cords send innumerable branches to the ovules scattered 

 over the placental plates. It was noted that the number of vascular bundles in 

 the pedicel did not always correspond to the number of placental bundles in the 

 fruit wall. The vascular system as a rule becomes stable just below the flower 

 head and remains unchanged until the insertion of the first leaf-trace, but occasion- 

 ally forking or fusion occurs, resulting either in an increase or a diminution of the 

 number of the bundles from that seen at a higher level. The observations made 

 on the mode of insertion of the leaf-traces agree on the whole with those of Leger, 

 though more variations were met with. Just before the entry of the first leaf- 

 trace the bundles of the stem lose their definite arrangement in three concentric 

 circles and form a more or less undulating ring. The bundles of the leaf on 

 approaching its base fuse into three or more cords. Where the system consists of 

 three cords (fig. 12, a-f) the main central bundle forks and gives off anterior strands, 

 while the- adjacent stem bundles fuse laterally. With this compound bundle the 

 anterior strands unite. This composite mass then splits, and the four bundles of 

 the leaf-trace enter by the gap so created. Finally, the two branches, arising by 

 the forking of the median leaf-trace bundle, unite with the lateral bundles of the 

 same leaf-trace and run down the axis as two independent cords. After a time 

 these bundles split into smaller strands and the normal arrangement of the vascular 

 system becomes established. 



Leger states that anterior strands are given off from all the leaf-trace bundles. 

 That is not so ; further, in some cases no anterior strands are formed at all. 

 Numerous variations are met with. Thus in one instance where no anterior strands 

 were formed the entire leaf-trace did not enter by the main gap produced by the 

 forking of the compound stem bundle ; in another case a special layer of parenchyma 

 enclosed both the leaf-trace and the vascular system of the axillary branch. The 

 leaf-traces at the base of the stem are simpler, for in some only one bundle enters 

 the foliar gap. The smaller anterior strands, instead of fusing into' an arc and 

 uniting with the axis system, may remain distinct and persist in this condition for 

 a considerable distance. The vascular tissue of the axillary branch is inserted one- 

 half on either side of the leaf-trace, just as the latter enters the stem. 



The stem anatomy of other species of Papaver need be described only in so 

 far as it shows differences from that of P. Rhceas. 



Papaver dubium, L. — The layer beneath the epidermis usually forms a well- 

 defined exodermis in Leger's sense of the term, but not infrequently it is indis- 



