October 22, 1915] 



SCIENCE 



579 



This paper makes no pretense of being an 

 exhaustive treatment of the subject under 

 consideration. Its main object has been to 

 point out as briefly as possible the danger of 

 assuming that the coefficient of correlation is 

 necessarily a satisfactory measure of all forms 

 of relationship between two variable quantities, 

 and at the same time to suggest a method of 

 attack for determining in what way a partic- 

 ular relationship depends on the value of this 

 coefficient. Charles N. Moore 



University of Qncinnati 



an aberrant ecological form of unto 

 complanatus dillwyn 



The variety of Unio complanatus Dillw. 

 which is here described was found at Songo 

 Pond, about three miles south of Bethel, Me. 

 The specimens from which it is described were 

 collected in August, 1913. The pond is a 

 headwater of the Crooked Eiver, one of the 

 larger tributaries of the Presumpscot. It lies 

 in a glacial scoop in alluvial sand, and 

 is fed by springs mainly. A small brook a 

 mile long enters it also. The country rock is 

 a granitic gneiss of the eastern range of 

 Montalban gneisses, and the intrusive granites 

 scattered here and there are of the same min- 

 eralogy. There is no limy rock in any form 

 within many miles, a fact which will account 

 for the peculiar structure of the shell. The 

 specimens were picked up on a very gently 

 sloping beach of round-grained sand, along the 

 western shore of the pond, and in about two 

 feet of water. The pond is about a mile and a 

 quarter long, from north to south, and aver- 

 ages a quarter of a mile in width. 



So far as I can determine, the soft parts of 

 the animal are in every way normal for the 

 species. The aberrancy occurs in the valves, 

 and is in structure and in shape. 



The largest of my specimens, and the largest 

 I have seen in the course of eight summers' pick- 

 ing, measures two and three quarters by one 

 and a half inches over all. The greatest thick- 

 ness, from umbo to umbo, is three quarters of 

 an inch. The following features are normal: 

 hinge size and place, umbo size, place and 

 shape, lateral and pseudocardinal teeth size 

 and shape, scars, pallial line, and sculpture. 



Epidermis is of normal color, but thicker than 

 usual, and overlaps the edge of the hard part 

 of the shell up to 3/32 of an inch, being most 

 extended at the siphonal region and along the 

 anterior part of the ventral edge in many speci- 

 mens. 



The shape of the shell is almost identical 

 with that of Anodonia marginaia Say, being 

 roughly rhomboidal. It does not resemble the 

 specimens of Unio complanatus from other 

 regions in the American Museum at New 

 York, in this respect. Prom the posterior end 

 of the hinge, the dorsal edge slopes ventrally, 

 straight,, at an angle between 35 and 40 de- 

 grees from the line of the hinge. This por- 

 tion of the edge is nearly straight and about 

 as long as the hinge. It rounds oif into the 

 small semicircle of the posterior end. In ma- 

 ture specimens there may be a slight flatten- 

 ing of the posterior end at the point where the 

 mantle forms a pair of siphons by its folding 

 and coherence, but this is not constant and I 

 find it only in the largest specimens. The 

 ventral edge is not a uniform curve, but ap- 

 proaches more or less to three straight lines, 

 equal in length, each making an angle of 

 about ten degrees with the line continuing the 

 edge beyond it. The anterior end has the 

 usual graceful elliptical outline, forming a 

 large curve from hinge to ventral edge. 



There are no rays visible on any of my 

 specimens. 



The most peculiar feature of the shell is the 

 exceedingly small amount of mineral matter 

 in it. "When fresh the shells are horny and 

 somewhat flexible, not unlike two layers of 

 parchment pasted together, in texture. Alco- 

 holic material and fresh are alike easily cut 

 with a small shears, and there is no cracking. 

 The thin nacreous layer breaks into small 

 angular chunks, which adhere to the epidermis. 

 I found only the faintest traces of a prismatic 

 layer, in the largest specimens. Smaller ones 

 fail entirely to show it. In my largest speci- 

 mens there is at the umbo a larger amount of 

 mineral matter, but even here it is hardly more 

 in amount than at the margin in the normal 

 shell of this species. The epidermis seems to 

 me to be nearly twice as thick as in the normal 

 type. In many specimens I found grains of 



