250 



SCIENCE 



[N. S. Vol. XXVIII. No. 712 



it remains incomprehensible why this decision 

 of the method of making up a deficiency, 

 which can be made only by the individual 

 teacher in the individual case, should deter- 

 mine a difference of grade. The grade to be 

 recorded on the books of the institution should 

 signify the student's rank and nothing else. 



We now have before us this entirely prac- 

 tical question: If an institution adopts a sys- 

 tem of grading like the one proposed, in which 

 3 per cent, are called excellent, 22 per cent, 

 superior, 50 per cent, medium, 22 per cent, 

 inferior and 3 per cent, failure, how can the 

 individual teacher, who is perhaps in charge 

 of a class of only five or eight students, com- 

 ply with the system? There is only one an- 

 swer to this question: He must work out his 

 method of grading for himself on the basis of 

 his individual experience with the students. 

 But he should be given one kind of aid by the 

 institution which he serves. The institution 

 should publish annually a statistical table 

 showing how each teacher has graded all his 

 students the last year and the last five years, 

 so that each teacher can inform himself easily 

 as to whether he has graded his students in 

 accordance with the system adopted by the 

 institution or has unconsciously applied an 

 arbitrary standard of his own and thus intro- 

 duced confusion into the system. There can 

 be little doubt that this would soon result in 

 a great uniformity of grading, and inequali- 

 ties of the size described would be impossible, 

 to the satisfaction of both faculty and stu- 

 dents. 



One problem is still left. How should the 

 ability of the five groups of students be repre- 

 sented in order to compute the claims of vari- 

 ous students for honors which are to be given 

 to those having the highest rank of a whole 

 student body. The University of Missouri 

 prescribes for this purpose that the first grade 

 be represented by 95, the second by 85, the third 

 by Y5 and the fourth by 65. These values 

 are so arbitrarily chosen that any one can see 

 that no scientific influence has been effective 

 grading on the probability curve, as we have 

 tried to do, we are able to give a reasonable 

 answer to the present question. In Fig. 2 



the ability of the average medium student is 

 found at the point where the abscissa is 0. 

 The ability of the average superior student is 

 found near -j- 1, that of the inferior student 

 near — 1, The ability of the average excel- 

 lent student is found near + 2, that of the 

 average failure student near — 2. All these 

 differences of ability are represented by steps 

 which are about equal. To avoid negative 

 values, it would, therefore, be the simplest 

 method to represent the different grades agreed 

 on by the numerical values 5, 4, 3, 2 and 1, 

 and to multiply these values by the number 

 of hours of work for which each grade has 

 been received. The students whose totals are 

 highest — making allowance for the probable 

 error, which is about .04, if the total number 

 of grades recorded during the college course 

 is about 40 — ^have then the best claims for the 

 honor as far as scholarship is concerned. 

 Max Meyer 

 Univeesitt op Missouei 



a new color variety of the guinea-pig* 

 Experimental studies made in recent years 

 show that color inheritance in mammals is a 

 matter of considerable complexity, but not 

 beyond the possibility of analysis. The more 

 carefully the matter is studied, the clearer 

 does the fact become that color inheritance, 

 in all its phases, conforms with Mendel's law 

 of heredity. The seemingly complicated re- 

 sults are due to multiplicity of factors con- 

 cerned in the production of those results. If 

 we confine our attention to one factor at a 

 time, we find that its behavior is strictly and 

 simply Mendelian. Each factor is either pres- 

 ent or absent and in general the presence of a 

 factor is dominant over its absence. It is 

 only when two or more independent factors 

 are simultaneously concerned that complica- 

 tions arise. Thus two simple factors acting 

 simultaneously may produce a result different 

 from that of either factor by itself. 



In the issues of Science for January 25, 

 1907, and for August 30, 1907, I have advo- 

 cated the view (first advanced concerning mice 



' Published by permission of tbe Carnegie In- 

 stitution of Washington. 



