1917] Taxonomic Value of A ntennal Segments 265 



Marlatt (1) stated that he had the gravest doubts as to the 

 value of the measurements of antennal segments in the deter- 

 mination of species, because the antennal formulae varied so 

 greatly even in the same individual. Pergande (\), Smith (3), 

 and undoubtedly many others have been misled into thinking 

 similarly because of the apparent discrepancies which arise in 

 the antennal formulae of any particular species, but which, if 

 carefully worked out by the curve method, would present a 

 definite working basis for that species. Quoting Cockerell (4) : 

 "It does not follow, as some uncritically assume, that antennal 

 measurements are useless" just because the antennal formulae 

 are so variable. 



Due to the fact that no published directions for making the 

 curves were known to the writer at the time this investigation 

 was started, he proceeded along lines of his own initiative and 

 used a system for the graphic representation of the antennal 

 segments which not only shows the general trend of the individ- 

 ual curves, but also shows the mean curve for all the measure- 

 ments of normal antennae in any particular species. 



Using millimeter graph paper, the writer laid off in succes- 

 sion, working downward on the vertical axis or axis of ordinates, 

 distances of ten millimeters. Each unit of ten millimeters was 

 let represent one antennal segment. It is needless to mention 

 that this unit was an arbitrary one, for the writer might just 

 as well, except for lack of space, have taken larger units. On 

 the horizontal axis, or axis of abscissis, the lengths of the 

 successive antennal segments were plotted, one millimeter 

 being used to represent a length of one micron. The writer 

 could have taken a different unit, but for the sake of simplicity 

 and compactness, the above-mentioned one was chosen. These 

 units have been used in all the accompanying curves, and con- 

 sequently they are uniform and comparative. Connecting the 

 points in succession along the direction of the vertical axis 

 produced the graphic figures. In all cases under observation, 

 these curves are distinctive and characteristic of each species 

 the writer has studied, as a comparison of the figures will show. 

 All figures have been photographically reduced one-half. 



After having followed this system for producing the antennal 

 curves, the writer learned, through the United States Bureau of 

 Entomolog^^ that Brain (Q) had published a few years pre- 

 viously the graphic representations of the antennal segments 



