28 ANATOMY OF THE GYMNOSPERMS 



The variations thus indicated are usually accompanied by a 

 more or less marked alteration in the relative volumes of the 

 spring and summer woods, in consequence of which the same 

 tree may show regions of coarse-grained and fine-grained wood. 

 Plate i exhibits the characteristic features of such fine-grained 

 wood, and a comparison with plate 46 will serve to emphasize 

 the deviation from the usual type of structure. 



Eccentricity of the growth ring is a feature more or less com- 

 mon to all woods (31, 5 13), and it is determined by external con- 

 ditions of light and warmth (22, 41 et seq.}, but ordinarily such 

 variations are not sufficiently marked to merit comment. In the 

 genus Juniperus, however, eccentricity is developed to an un- 

 usual extent, and it serves as a more or less distinguishing fea- 

 ture. From the photograph presented in plate 2 it will be seen 

 that the rings not infrequently coalesce on one side, while gain- 

 ing great prominence as separate structures on the opposite side. 

 A like eccentricity characterizes the genus Taxus. 



We are then to conclude that growth rings are a normal and 

 constant feature in the stem structure of the Coniferales as a 

 whole, and the same is also true of the Gingkoales so far as we 

 know them through both fossil and recent examples of the genus 

 Gingko. But the same law does not apply to the Cordaitales as 

 a whole, and in this may be found one of the leading distinctions 

 between these two groups. This is especially true of the genus 

 Cordaites in which the " growth rings when present are obscure, 

 rarely somewhat conspicuous," and even in the latter case they 

 appear simply as regions of somewhat unequally variable density, 

 dependent upon regional changes in the thickness of the tra- 

 cheid walls and the volume of the lumen, the one region merging 

 into the other by somewhat gradual transitions and always 

 without that sharply defined alteration of structure so charac- 

 teristic of the growth rings in the Coniferales (plate i). In 

 existing Araucarias the " growth rings are not determinate, or 

 at most poorly defined," though De Bary (13, 513) cites the case 

 of a specimen of A. excelsa grown in the open ground, which 

 showed sixteen sharply defined growth rings, and Kraus has 



