Studies reporting predictions from assessments of infant recognition memory to later developmental performance and IQ are critiqued. Inelegancies in design, procedure, and analyses are noted. While this approach may have potential, its utility for practical or clinical purposes is still not demonstrated. (Author/RD)