The assumption that intraspecific variance is low across brain regions has led many comparative studies to rely on small sample sizes to determine volume, neuron number, and soma size. If, however, there is significant intraspecific variation in brain measurements, sampling only one or two individuals may lead to biased or incorrect conclusions. Here, we examined intraspecific variation in two sensory brain regions, nucleus magnocellularis (NM) and nucleus rotundus (nRt), among a large sample of ruffed grouse (Bonasa umbellus). Using unbiased stereology, we quantified volumes, neuron numbers, and soma sizes in both NM and nRt. Overall, larger brains did not have larger volumes, more neurons, or larger neurons in either brain region. Larger nRt volumes had more neurons, but this was not true of NM, and we found no significant relationships between region volume and soma size. Our data, therefore, indicate that within species, larger brain regions do not necessarily have more neurons. We then used Monte Carlo simulations to estimate confidence intervals (CI) across a range of sample sizes for NM and nRt volume, neuron number, soma size, and neuronal density. As predicted, low sample sizes had large variation around the sample mean and this variation progressively narrowed with larger sample sizes. At a sample size of n = 4-7, the bounds of the CIs began to decrease greatly and ultimately stabilized at n = 12-13 at which point the CIs closely approximated that of the sample mean. Based on these results, we advise using larger sample sizes whenever possible, and small sample sizes (< 4) should be used cautiously in comparative neuroanatomical studies.