But a number of anomalies remain. The American substitution of /a/ for /e/ in /gray/ is not easily explained, nor is the substitution of /k/ for /c/ in /skeptic/ and /mollusk/, nor the retention of /e/ in /forego/, nor the unphonetic substitution of /s/ for /z/ in /fuse/, [Pg247] nor the persistence of the first /y/ in /pygmy/. Here we have plain vagaries, surviving in spite of attack by orthographers. Webster, in one of his earlier books, denounced the /k/ in /skeptic/ as "a mere pedantry," but later on he adopted it. In the same way /pygmy/, /gray/ and /mollusk/ have been attacked, but they still remain sound American. The English themselves have many more such illogical forms to account for. In the midst of the /our/-words they cling to a small number in /or/, among them, /stupor/. Moreover, they drop the /u/ in many derivatives, for example, in /arboreal/, /armory/, /clamorously/, /clangorous/, /odoriferous/, /humorist/, /laborious/ and /rigorism/. If it were dropped in all derivatives the rule would be easy to remember, but it is retained in some of them, for example, /colourable/, /favourite/, /misdemeanour/, /coloured/ and /labourer/. The derivatives of /honour/ exhibit the confusion clearly. /Honorary/, /honorarium/ and /honorific/ drop the /u/, but /honourable/ retains it. Furthermore, the English make a distinction between two senses of /rigor/. When used in its pathological sense (not only in the Latin form of /rigor mortis/, but as an English word) it drops the /u/; in all other senses it retains the /u/. The one American anomaly in this field is /Saviour/. In its theological sense it retains the /u/; but in that sense only. A sailor who saves his ship is its /savior/, not its /saviour/. § 3