Results of experimental transfer of rainbow smelt into lakes reclaimed by rotenone around 1960 in Maine were originally interpreted to cast doubt on the previously widely accepted hypothesis that there were two hereditarily different forms of rainbow smelt, one large and one small. Study of more recent data from some of the transplanted populations and reanalysis of the original data suggests different conclusions. The initial effect of introductions into a reclaimed lake may be accelerated and/or more prolonged growth which exceeds even interlake differences. This initial phase, however, is followed by a second phase, when the population reaches equilibrium and these effects subside. Data from phase two of the Little Concord-Shagg and Cold Stream-Coleback Lake transfers showed that the growth characteristics of the transplanted populations returned to those of the parental populations. Large differences in growth patterns were thus found only in the initial phase of the introduction. Meristic characters were little affected by transplanting.
Analysis of large specimens derived from a postulated second unofficial introduction into Coleback Lake showed that they also differed significantly, having both higher gill raker and vertebral counts than the smaller smelt. This was of interest as smelt vertebral and gill raker counts usually are inversely related; hence we do not equate these for the moment with the large form of smelt known elsewhere.
It is concluded that the initial interpretation of transfer experiments be delayed until conditions approaching equilibrium can be expected to exist. Further, our analysis of more recent lake transfer data has shown nothing to refute the hypothesis that there are at least two hereditarily different forms of smelt.