Astron.Astrophys. 425 (2004) 615-626 We examine the large sample of lambda Boo candidates collected in Table 1 of Gerbaldi et al. (2003) to see how many of them show composite spectra. Of the 132 lambda Boo candidates we identify 22 which definitely show composite spectra and 15 more for which there are good reasons to suspect a composite spectrum. The percentage of lambda Boo candidates with composite spectra is therefore > 17 and possibly considerably higher. For such stars the lambda Boo classification should be reconsidered taking into account the fact that their spectra are composite. We argue that some of the underabundances reported in the literature may simply be the result of the failure to consider the composite nature of the spectra. This leads to the legitimate suspicion that some, if not all, the lambda Boo candidates are not chemically peculiar at all. A thorough analysis of even a single one of the lambda Boo candidates with composite spectra, in which the composite nature of the spectrum is duly considered, which would demonstrate that the chemical peculiarities persist, would clear the doubt we presently have that the stars with composite spectra may not be lambda Boo at all.