( '>47 ) 



been necessary for 36 combinations. However, as each case is repeated 

 once, there are in reality 756 separate judgments. 



Just now we said it was necessary that never other vectors than 

 those about equally directed, which accordingly possess about the 

 same unit of length, should be mutually compared. If we keep this in 

 view, the following constellations are identical within a limit of I ' , : 



isoamylaeetate and nitrobenzol 

 valerian acid and scatol 



isoamylaeetate and guajacol 

 nitrobenzol and guajacol 

 muscon and scatol 



aethvlbisulfide and valerian arid 



each four t imp- 



each three times 



isoamylaeetate and terpineol 



isoamylaeetate and aethvlbisulfide 



nitrobenzol and terpineol 



terpineol and guajacol each once 



muscon and aethylbisulfide 



muscon and valerian acid 



aethylbisulfide and scatol 



together 27 constellations '). 



Some of these 27 constellations give rise to reciprocity, in such a 

 way, that the vet-tors which are identical to two previously deter- 

 mined vectors, make the latter identical, when they themselves are 

 previously determined in their mutual situation. This occurs in : 



isoamylaeetate <■— .muscon, — » i isoamylaeetate 



and with regard to and with regard to and 



terpineol — *• ( scatol ) «- guajacol 



Though the coincidence of two odour-vectors, if considered with 

 respect to two other odour-vectors, is already very remarkable, a 

 coincidence of 3 vectors is still more interesting. This has been 

 realized in the following cases, if we extend the limit of error to 2° „ : 



') This figure rises to 42, if besides complete coincidences also coincidences 

 with reciprocal values are taken into consideration. 



