v; PREFACE. 



The classification presented is in accordance with recent 

 studies and the conviction that certain of the lower so-called 

 ••orders" of ine ts, such as the " < >rthoptera." "pBeudo- 

 neuroptera," are heterogeneous, unnatural groups, which 

 for the sake of clearness and truth to nature should be 

 broken up into distinct orders. The class of insects, ti. 

 fore., is divided into sixteen orders insi t eight, as may 



-■•en in tabular form on p. 56, and the ue 

 of orders has been reversed, the book beginning with the 

 lowest, the wingless insects, and ending wi.th the big": 

 the Hymenoptera. This ordei _ - with the probable 

 mode of evolution of the class, and with the _ 



-OQ of ins "-. - - we know it: insects like cock- 



roaches., grasshop: _ the first to appear, those 



with a metamorpl - -. as nenroptera, beetles, flics, moths, 

 ants, and I -. succeeding them. 



In 1SG3 the author proposed a new classification of i 

 placing the Hymenoptera at the head of the ii ies, 



the Ooleoptera having, because, perhaps, from being the 

 favorites of collectors, been - g ted this position. Since 

 that time it has been gratifying to see that, at the present 

 time, not only in the United Si -. but in England and on 

 the Continent, the Hymenoptera by geneml consent crown 

 the summit of tl 



The present scheme of classification was in part worked 

 out by the author (contrary to his early convictions or preju- 

 -I and published in 1883, when ten orders instead <>f 

 eight were adopted; while it was remarked that the three 

 groups (i.e., white ai . tin- may-flies, and the dragon- 



mposing the •• Pseudoneuroptera" might hereafter 

 he i- _ i as entitled to the rank of orders. It should 



be l>orne in mind that some of the leading entomolog 

 as Westwood and others, had for many years regarded the 

 ear-wigs, the « — » . i -. tin- Thripida\ and the flea- 



representing distinct orders. While we wire considering it 

 a debatable question whether these important types had not 

 unwarrantahly "lumped" with the older Linnaan 



