Length frequency distributions obtained from the FIMS of the west coast rock lobster constitute one of data inputs into the assessment as well as an estimate of the percentage of females. The length composition of the catch for each station and leg is obtained from measurements of a sub-sample of the catch, or sometimes the whole catch is measured. Nominal length frequencies (un-weighted frequencies pooled across stations) for males and females have been used previously as inputs to the assessment model of the rock lobster resource. The same is the case for the percentage of females. If the sampling of rock lobsters for measurement represents a random sample from a station and the sample is representative of the size composition of rock lobsters, then no weighting of the data is necessary. However if this is not the case, then the length frequencies should be weighted before pooling. This paper investigates length frequency distributions of rock lobster by sex obtained in various ways. A method is suggested that should be used in constructing length frequency distributions to be used as an input to the assessment model. Annual estimates of the female percentage of rock lobster, weighted in a similar way to that of the length frequency distributions, are also given. In the FIMS database there are several records that have the total number of rock lobsters measured greater than the total number of rock lobsters caught. Some of these records have been corrected, but for some this was not possible as the error also occurred in the original records. For the analyses presented in this paper, this error has been ignored and the values given as the total number of measured rock lobsters have been taken as being correct.