The frequency distributions of sizes of "clouds" and "clumps" within clouds are significantly flatter for extinction surveys than for CO spectral line surveys, even for comparable size ranges. A possible explanation is the blocking of extinction clouds by larger foreground clouds (occlusion), which should not affect spectral line surveys much because clouds are resolved in velocity space along a given line of sight. We present a simple derivation of the relation between the true and occluded size distributions, assuming that clouds are uniformly distributed in space or are all at about the same distance. Because the occlusion is dominated by the largest clouds, we find that occlusion does not affect the measured size distribution, except for sizes comparable to the largest size, implying that occlusion is not responsible for the discrepancy if the range in sizes of the samples is large. However, we find that the range in sizes for many of the published observed samples is actually quite small, which suggests that occlusion does affect the extinction sample and/or that the discrepancy could arise from the different operational definitions and selection effects involved in the two samples. We suggest that studies of isolated small dark clouds and of IRAS cloud cores support the former interpretation.