havior were more consistent than those seen with meth ylphen idate , which differs from placebo only at marginal levels (p = 0.056) at this stage of the study . As in other studies , fenfluramine caused drow­ siness in some children (although not severe in any case), and the differences approached significance both in comparison with placeb o (p = 0 .052) and meth ylphen idate (p = 0.052). We observed a non­ significant trend for fenfluramine to cause anorexia, but so far, this is less apparent than the differences between methylphenidate and placebo (p < 0.10 ). We have also a see n a nonsignificant tenden cy for less akathisia with fenflur amine than with placebo (p = 0 .06 8), which may be as a result of a mild anxiol ytic effect or may be secondary to the reduction in activity level. One heretofore unreported side effect is a very vivid report of a 7-year-old girl who developed a peculi ar body odor and halitosis that co incided exactly with the times of fen­ fluramine treatment. We observed no increases in insomnia, dizziness , or diarrhea with fenfluramine , although these are sometimes report ed to occ ur with the drug (Aman and Kern , 1989). We also looked at heart rate, blood pressu re, and body weight in these children. Analysi s of varian ce indic ated that all three cardio­ vascular measures were significantly affected by the medications . Fen­ fluramine was associ ated with reductions in eac h, but all sig nifica nt comparisons were with respect to methylphenidate , which tended to elevate heart rate and blood pressure . As in most studies, fenfluramin e produced significant weight losses in these children (p = 0. 01) . Our subjects lost about 1.8 pounds (0.8 2 kg) relative to their placebo weights during the 4 weeks the y received medic ation . Finally, two groups of sympt oms that do not appear on the DOTES Scale occurred fairly often in our sample. Seven of our subjects were reported to have difficulties wi th en uresis and encopresis, and another four demon strated stereotyp y and self-in jurious behavior. Enure sis and encopresis were lowest when the childre n were taking methylphen i­ date , appearing to worsen both with placebo (p = 0 .068) and fen­ fluramin e (p = 0 .068 ) . The emergence of these symptoms (usually in childr en who had previously been maintained on methylphenidate) has probably caused more distre ss than any other side effect in this study . Furthermore , it is a well-established fact that moderate-t o-high doses of psychostimul ants can induce stereotypic behavior in labora­ tory animal s (Randrup and Munkvad, 1974 ), and sterotypies may occur de novo (or existing stereo typies may be exacerbated) at lower doses of stimulants in children with developmental disabilities (A man, 1982). Both enure sis/encopresis and stereotypy/self-injury are obviousl y of rele vance to researchers and cli nicians working with de velopmentally delayed clinical popul ations. Other researchers working with devel­ opmentally delayed children may wish to add these symptoms to their inventory of side effec ts if they are using the DOTES to monitor side effects. Furthe rmore , if the DOTES is revi sed in the future , its de­ velopers should seriously consider add ing these and other relevant items to document such potential side effects .