236 E. B. TITCHENER— ETHNOLOGICAL TESTS OF SENSATION. 



In conclusion I offer a tentative suggestion as regards the gen- 

 eral conduct of field-tests ; it is always dangerous to be positive, 

 but I take the risk. There is, I understand, a present tendency 

 among those interested in mental tests to break away from tests of 

 the " all or none " type and to substitute for them a set of tests which 

 permits of fractional grading. The "all or none" tests, it is argued, 

 cannot be applied to a long series of subjects, whereas tests which 

 may be rated for part-performance have a practically unlimited 

 range of application. I suppose that both kinds of test will be re- 

 tained, each in what turns out to be its proper sphere; and I am 

 disposed to think that, for anthropological purposes, the " all or 

 none " type should, at first and on the whole, be preferred. Every- 

 one who has worked with Hering's instruments must have been 

 struck by the fact that they serve admirably for their one predestined 

 experiment but that they can with great difficulty, if at all, be adapted 

 to other uses. The demands of undergraduate teaching have led 

 most of us, perhaps, consciously or unconsciously to favor instru- 

 ments of a more flexible, more variously usable sort. Yet it may 

 be that, for the primitive subject, tests of Hering's kind are, at 

 least in the beginning, the more desirable. I wonder if a large 

 number of testlets, each one sharply cut to a particular purpose, 

 might not be better than tests which require serial or repeated obser- 

 vation; and if the single-value result might not lend itself to mathe- 

 matical treatment better than the somewhat arbitrarily chosen " rep- 

 resentative " value of a test-series. 



