We hypothesise that of the two processes underlying declarative memory, recollection is impaired in high-functioning autism (HFA) whereas recollection and familiarity are impaired in low-functioning autism (LFA). Testing these hypotheses necessitates assessing recollection and familiarity separately. However, this is difficult, because both processes contribute to performance on standard memory tests. Moreover, tests must be suitable for use with young or intellectually disabled participants. This study aimed to develop tests of recollection and familiarity separately, and to make preliminary tests of our hypotheses. We developed a temporal source memory task to assess recollection in LFA, and a shape recognition task to assess familiarity and an action recall task assessing recollection in HFA. The methods and implications of the results are discussed.