This study examined the hypothesis of superior quantification abilities of persons with high functioning autism (HFA). Fourteen HFA individuals (mean age: 15 years) individually matched with 14 typically developing (TD) participants (gender, chronological age, full-scale IQ) were asked to quantify as accurately and quickly as possible numerosities, represented by the number of squares (2-9) presented in random configurations. In addition, the visual angles of stimuli presentation were manipulated in order to induce a local (large visual angle) and a global (small visual angle) bias on participants' quantification performance (accuracy and naming time). Findings revealed no effect of local and global bias of stimuli presentation in the two groups' performance, and no superior quantification abilities in HFA participants. However, analyses of the naming time slopes for identification by HFA participants of small consecutive numerosities (2-5), suggested their use of counting processes instead of subitizing (or immediate apprehension of numerosities) as in TD participants. Possible explanations for these results are discussed with reference to models of locally-oriented information processing in autism.