1909] COULTER—EVOLUTION AMONG GYMNOSPERMS 85 
genetic connection of these types of cylinder, the siphonostele is the 
type that was carried forward in the evolution of gymnosperms. 
This siphonostele was made up of collateral mesarch bundles, 
developed secondary wood composed of tracheids, and the bundles 
of all its peripheral connections were concentric or at least mesarch. 
Among the gymnosperms the universal tendency was to eliminate 
the centripetal xylem, a tendency carried forward from the ferns, 
until the collateral mesarch bundles of the central cylinder became 
collateral endarch; and more gradually the bundles of the peripheral 
regions became collateral mesarch and finally collateral endarch. 
So early was this accomplished for the central cylinder that a collateral 
endarch cylinder is a feature of gymnosperms in general. From 
what has been said as to the variable rate of change among the mem- 
bers of a great group, it would be expected that mesarch and even 
concentric bundles might be found in peripheral parts of certain 
species of genera whose allies had completely eliminated centripetal 
xylem, or might occur occasionally in any species or genus. It is 
an interesting fact that centripetal xylem appears to linger longest 
in the cotyledons; and the number of gymnosperms in which it is 
own to occur in this organ, regularly or occasionally, is increasing 
rapidly. ; 
In the development of secondary wood, the general tendency 
among gymnosperms is to increase it in amount, so that a thick 
vascular cylinder is built up by the primary cambium. This tend- 
ency is apparent among the Cordaitales, but it reaches its most con- 
spicuous result in the Coniferales and Ginkgoales. This changed 
also the general topography of the stem, both pith and cortex being 
much reduced in relative amount. The Bennettitales and Cycadales 
responded feebly if at all to this tendency, one of their features being 
the retention of the general stem structure of the Cycadofilicales. 
In these groups the primary cambium is either short-lived or functions 
very slowly, and in some forms secondary cambium produces cortical 
bundles; but the formation of secondary wood never prevents the 
formation of a large pith and an extensive cortex. It is in these 
Sroups also that the concentric and mesarch bundles are most com- 
mon in the peripheral members, being found somewhere in all 
forms; while among the Coniferales the centripetal xylem has been 
