NOVEMBEB 12, 1909] 



SCIENCE 



691 



grained quartz sands, cemented by calcium 

 carbonate. So far as observed they do not 

 vary appreciably in width through vertical 

 range. Two joint systems, one nearly hori- 

 zontal, the other vertical, have cut these dikes 

 in such a manner as to suggest masonry walls, 

 i. e., they are composed of oblong blocks in 

 horizontal layers. 



Certain facts may be noted, however, which 

 preclude this view. In a photograph at hand 

 exposing a portion of the dike near Rockwall, 

 it may be seen that many of the vertical joints 

 occur above each other, i. e., they are not 

 broken, which condition would not exist in a 

 wall constructed by hand. It may also be 

 noted that the curve to the upper surface of 

 one block exactly fits the curve on the under 

 surface of the next block above, which leads 

 to the same conclusion. The weathered sands 

 between the joints, stained with iron oxide, 

 have been mistaken for mortar. 



To define accurately the steps which have 

 taken place in the forming of these dikes is 

 not as easy as to recognize the nature of the 

 phenomenon. They may have originated in 

 several ways. The sands may have come from 

 above or from below. The cracks may be due 

 to drying or to earth movements. The writer 

 was not able to decide the direction from 

 which the sands entered. Inasmuch as cir- 

 culating waters have passed for long periods 

 through the sands, dissolving and redissolving 

 the cement between the grains, the original 

 position of the latter can not be postulated. 

 At present they show no signs of bedding. On 

 breaking blocks, what might be called a stalag- 

 titic fracture is obtained, i. e., cylindrical or 

 tubular forms arranged in vertical position. 

 As has been pointed out, this may well be 

 secondary structure induced by circulating 

 water. 



The limey muds were probably deposited in 

 very clear quiet waters. A slight elevation of 

 the sea or an increased supply of material 

 from the land may have altered deposition and 

 spread fine sands upon the muds. Cracks 

 formed by earthquakes may have permitted 

 unconsolidated sand to enter as a filling. 

 Again, the muds may have undergone a dry- 



ing-out process since their elevation above the 

 sea, cracks may have formed from this cause, 

 and overlying sandy layers aided by perco- 

 lating waters served to supply material where- 

 with to fill them. 



The joints may be ascribed to forces arising 

 from slight warping of the earth's surface, 

 acting on hard vertical masses imbedded in 

 relatively plastic strata. 



It is fair to say in conclusion that the be- 

 lievers in the theory which ascribes the origin 

 of these dikes to prehistoric men are in the 

 minority in the locality itself. 



Sidney Paige 



apogamy in cenothera 



There seemed at one time a possibility that 

 the phenomena of mutation in (Enothera 

 LamarcTciana might be associated with a con- 

 dition of apogamy in that species. A survey 

 of the hereditary behavior, however, and par- 

 ticularly of the results of certain crosses be- 

 tween the mutants and 0. Lamarckiana, and 

 also among the different mutants themselves, 

 soon made it apparent that such a condition 

 could not be of high frequency at any rate, in 

 the parent form or in such mutants as 0. 

 rubrinervis and 0. nanella. The results of 

 crosses between 0. Lamarckiana or certain of 

 its mutants, and such wild species as 0. hien- 

 nis, also could only be explained by assuming 

 that fertilization had taken place uniformly 

 in the ordinary way, and often the resulting 

 hybrids show the predominating influence of 

 the pollen parent. 



But while it seems highly improbable that 

 apogamy in 0. Lamarckiana is concerned in 

 the origin of the mutants, yet, as I shall pro- 

 ceed to show, there is some very good evidence 

 that one at least of these mutants is itself 

 apogamous, though only in a small percentage 

 of cases. 



(Enotliera lata is well known to be sterile in 

 its anthers, so that self-fertilization has never 

 been effected. MacDougal' has reported that 

 the form closely resembling 0. lata, from near 

 Liverpool, England, can be self-fertilized, and 



' " Mutations, Variations and Relationships of 

 the Oenotheras," Carnegie Inst., Pub. 81, p. 15, 

 1907. 



