THE FLORA. 



93 



locality for a single species beyond the state- 

 ment that they were collected at Golden and 

 came from the Laramie. It was ascertained 

 later from Arthur Lakes, who collected this 

 material, that so far as he could recall all came 

 from Table Mountain or Green Mountain, 

 and that "none of them came from the prox- 

 imity of the lower coal measures." This 

 material is the property of Harvard University 

 and has not been reviewed in the present con- 

 nection, but I saw it casually some years ago 

 and do not recall any specimens not preserved 

 on the matrix characteristic of the Denver 

 formation at Golden. 



The difficulty, not to say danger, of relying 

 implicitly on the matrix to separate collections 

 of questioned stratigraphic position is well 

 illustrated in the attempts made to locate the 

 early collections from Coal Creek, east of 

 Denver. (See p. 90.) This material is pre- 

 served on a very soft carbonaceous sandstone 

 and was supposed to have come either from 

 the extreme upper part of the Laramie or from 

 the Arapahoe formation. Recent information 

 derived from a study of the log of a deep well 

 makes it probable that these Coal Creek plants 

 did not come from either the Laramie or the 

 Arapahoe, but from the Denver formation. 

 The study of the matrix of these plants did 

 not and could not settle the horizon with', 

 certainty. 



Another source of difficulty in dealing with' 

 the older collections has arisen in the following, 

 manner: It was Lesquereux's custom to pub- ; 

 lish preliminary accounts, in the annual re- 

 ports of the Hayderl Survey and elsewhere, of 

 the flora of various localities and formations. 

 These included new species more or less ade- 

 quately described and previously known spe- 

 cies, some of them European, but when the 

 formal presentation of the floras was made, 

 as in "The Tertiary flora," or "The Cretaceous 

 and Tertiary floras," certain of these species; 

 were merged with other forms and no syno- 

 nymic record was made of such combinations.; 

 It is thus practically impossible now to deter J , 

 mine the disposition that was made in such 

 cases, and this explains why forms once re- 

 ported as present in a flora are no longer 

 acceptedj or if admitted are so carefully quali- 

 fied. A case in point is offered by Cercis 

 eocenica Lesquereux. In 1873 this species was 

 85344—22 7 



named and imperfectly described from "Erie 

 mines, Boulder Valley," 7 but so far as known 

 it was not afterward alluded to by Lesque- 

 reux. It was never figured nor adequately 

 characterized, nor' is there known to be a 

 specimen ■ representing it. It was probably 

 united with some other species, but if so there 

 is no record of such transfer. 



Another example is furnished by Quercus 

 lyelli Heer, a European Tertiary form listed 

 by Lesquereux 8 in his first account of the 

 plants from Marshall, Colo. It was not men- 

 tioned again in this connection and has not 

 been since identified from that locality. 



Still another series of errors has come from 

 the obvious mislabeling of localities. Under 

 the system at present employed a locality 

 number is placed on each specimen, and it is 

 practically impossible to give a wrong locality, 

 but in the earlier days a single detached label 

 was all that accompanied a whole collection, 

 which might comprise dozens or even hundreds 

 of specimens. When such collections became 

 mixed, as there is undoubted evidence to show 

 that they sometimes did, the only possible way 

 in which the forms can be allocated is by the 

 matrix. If this is sufficiently distinct to be 

 characteristic, the specimens can be separated 

 with a fair degree of certainty, as, for example, 

 when a specimen labeled "Table Mountain, 

 Golden" — which should be in the Denver 

 formation — is found preserved on the whitish 

 arkosic sandstone known to belong to the 

 Laramie of the region, or when a specimen 

 preserved on the andesitic matrix character- 

 istic of the Denver formation is labeled " Black 

 Buttes, Wyo.," where there is a somewhat 

 similar appearing yellowish sandstone, the 

 error can be detected. When the matrix is 

 the same or practically the same, the case is 

 hopeless, and this is undoubtedly the cause for 

 much erroneous distribution of species, both 

 areal and vertical. 



I have taken the trouble to explain at length 

 certain of the difficulties encountered in evalu- 

 ating the original collections. Many toilsome 

 hours have been spent in trying to secure a 

 rational treatment of this material, but even 

 so, it is probably too much to assume that the 



',U. S. Geol. and Geog. Survey Terr. Rept. for 1872, p. 384, 1873. 

 - B Lesctuereax, Leo, [Notes on fossil plants from Bock Creek, etc.]: Am. 

 Jour. Sci., 2d ser., vol. 45, p. 208, 1868. 



