The Forest Insect Survey at the Forest Biology Laboratory, Winnipeg, has been compiling records for a number of years on the percentage of cocoons of the larch sawfly, Pristiphora erichsonii (Htg.), parasitized by the tachinid Bessa harveyi Tnsd. Sawfly cocoons were collected each fall from the soil in infested stands, and those containing living larvae were dissected to determine the percentage of parasitism by B. harveyi. These estimates have been used to provide an index of parasitism (Lejeune and Hildahl, 1954), but are of limited value for a number of reasons: (1) estimates can be expressed only as percentage of sound cocoons parasitized; (2) total parasitism cannot be estimated because a portion of the parasites emerge from the cocoons before collection; and (3) estimates of parasitism may not be representative of the stand because there is a tendency to collect cocoons where they are easiest to find; hence all the cocoons in a collection may be from one or two small areas. If the proportion of cocoons containing B. harveyi varies within a stand such collections may give unreliable estimates of parasitism.