A great many of his innovations, of course, failed to take root, and in the course of time he abandoned some of them himself. In his early "Essay on the Necessity, Advantage and Practicability of Reforming the Mode of Spelling" he advocated reforms which were already discarded by the time he published the first edition of his dictionary. Among them were the dropping of the silent letter in such words as /head/, /give/, /built/ and /realm/, making them /hed/, /giv/, /bilt/ and /relm/; the substitution of doubled vowels for decayed diphthongs in such words as /mean/, /zeal/ and /near/, making them /meen/, /zeel/ and /neer/; and the substitution of /sh/ for /ch/ in such French loan-words as /machine/ and /chevalier/, making them /masheen/ and /shevaleer/. He also declared for /stile/ in place of /style/, and for many other such changes, and then quietly abandoned them. The successive editions of his dictionary show still further concessions. /Croud/, /fether/, /groop/, /gillotin/, /iland/, /insted/, /leperd/, /soe/, /sut/, /steddy/, /thret/, /thred/, /thum/ and /wimmen/ appear only in the 1806 edition. In 1828 he went back to /crowd/, /feather/, /group/, /island/, /instead/, /leopard/, /sew/, /soot/, /steady/, /thread/, /threat/, /thumb/ and /women/, and changed /gillotin/ to /guillotin/. In addition, he restored the final /e/ in /determine/, /discipline/, /requisite/, /imagine/, etc. In 1838, revising his dictionary, he abandoned a good many spellings that had appeared in either the 1806 or the 1828 edition, notably /maiz/ for /maize/, [Pg252] /suveran/ for /sovereign/ and /guillotin/ for /guillotine/. But he stuck manfully to a number that were quite as revolutionary--for example, /aker/ for /acre/, /cag/ for /keg/, /grotesk/ for /grotesque/, /hainous/ for /heinous/, /porpess/ for /porpoise/ and /tung/ for /tongue/--and they did not begin to disappear until the edition of 1854, issued by other hands and eleven years after his death. Three of his favorites, /chimist/ for /chemist/, /neger/ for /negro/ and /zeber/ for /zebra/, are incidentally interesting as showing changes in American pronunciation. He abandoned /zeber/ in 1828, but remained faithful to /chimist/ and /neger/ to the last.