Formation of the Southern Indian Lake reservoir, northern Manitoba, added an estimated 5.4 × 105 t of Picea mariana (black spruce) needles to the lake. The breakdown of needles in the lake was measured by stringing needles on monofilament line, placing the strings into 3-mm mesh bags, and situating the bags along four shorelines representing high and low shoreline erosion rates (clay vs. bedrock) and wave exposures (highly exposed vs. protected). Sampling was done in phases of 0–41 and 328–384 d. Effects of excluding macroinvertebrates were tested by using needle strings placed in 50-μm mesh bags at the clay low exposure shoreline. Needle breakdown apparently occurred in two stages. Initially, weight losses were due primarily to leaching and microbial conditioning. Subsequently, weight losses were due primarily to feeding by macroinvertebrates. Processing coefficients (k) ranged from 0.0011∙d−1 for the 50-μm mesh bags to 0.0097∙d−1 for the 3-mm mesh bags at the bedrock high exposure shoreline. The presence of macroinvertebrates and exposure to waves enhanced needle breakdown. Macroinvertebrate colonization of needle strings occurred rapidly [Formula: see text] and was mainly by chironomid larvae. Phaenopsectra punctipes (Wied.) and Brillia flavifrons (joh.) larvae fed directly on needle tissue. Macroinvertebrates accounted for ~ 40% of total weight losses and, apparently, were active over the winter. Leaching and microbial conditioning each accounted for ~ 30% of total weight losses. Values of k for Southern Indian Lake most resembled those for conifer needle breakdown recorded in streams, indicating the significance of wave action in the lake. The breakdown of P. mariana needles in Southern Indian Lake appeared to be a significant source of carbon during the year following impoundment, and needles may have been an important habitat for macroinvertebrates within localized areas.