E. B. TITCHENER— ETHNOLOGICAL TESTS OF SENSATION. 235 



practised observers, and they thus provide a ready means of testing 

 any abbreviated test-method that may be proposed. Is it enough to 

 take 10 observations of a kind, and has the 8-in-io Hmen any definite 

 significance ?" Is it enough to take 5 observations, and has a 4-in-5 

 limen any vahie? These are questions not of pure but of appHed 

 mathematics ; they can be answered only in the Hght of comparative 

 data ; and the control-data are ready to hand in the laboratories. It 

 is useless to make tests in tlie field, and to repeat them later upon 

 civilized subjects, until we know whether the test-procedures are 

 themselves methodically reliable. So the laboratory may help on the 

 score of method. It may also help in other ways ; our discussion of 

 the perceptive forms in aesthesiometry applies, mutatis mutandis, 

 to more than one of the Torres Straits experiments. The analyses 

 of the laboratory show what the tests are really doing, what psycho- 

 logical level has been reached. The test of an optical illusion, for 

 example, may tell us nothing of the relative magnitude of the illusion 

 in the case of savage and civilized subjects, but may nevertheless 

 bring out the psychologically important fact that savage and civilized 

 approach the particular task set them in different ways, or come to 

 it in different attitudes of mind. 



I am not inviting the field-worker to fall between two stools ; I 

 am rather pleading for cooperation. The field-worker seeks to ob- 

 tain psychological data which shall enable him to rank a primitive 

 race in relation to the various strata of his own civilized community. 

 He knows, in a rough way, what can be done with a primitive popu- 

 lation ; the home-staying psychologist does not. The laboratory- 

 worker, on his side, knows that a good many of the tests commonly 

 employed are scientifically worthless ; yet he cannot be continually 

 playing the critic. Is it not a clear case for cooperation ? So far as 

 I know, we have to-day no approved aesthesiometric test that can be 

 carried into -the field. No: but if we settled, plainly and positively, 

 what it is that we want the test to tell us ; and if the field-worker 

 kept guard over complexity of technique and the laboratory- worker 

 over sources of error ; then a test would be forthcoming. 



''^ As regards the sesthesiometric test I have answered this question in 

 the negative. My own first series with the apposed compass-points gives 

 runs of 10 consecutive dual impressions with 8o per cent, correct judgments. 

 Unfortunately the whole run was not lo but i6. 



