A 1992 study of serotiny in lodgepole pine (Pinuscontorta Dougl. ex Loud. var. latifolia Engelm.) in Yellowstone National Park asked four questions: (i) are there morphological characteristics that can be used to estimate pre-fire proportion of serotinous trees in forests that burned in 1988?; (ii) at what spatial scale does percent serotinous trees vary across the landscape?; (iii) which environmental factors are correlated with serotiny?; and (iv) what is the relationship between prefire serotiny and postfire lodgepole pine seedling density? We first sampled cone characteristics in serotinous and nonserotinous trees along four 2950-m transects in unburned forests, and examined burned trees nearby. Results indicated that asymmetrical cones and an acute angle of cone attachment to the branch were reliable indicators of serotiny even in burned trees. We then sampled nine patches of lodgepole pine forest that had burned in 1988, and varied in size from 1–3600 ha. We sampled serotiny at varying intervals along two perpendicular transects that crossed in the center of each patch. At each sample point, the 12 nearest canopy lodgepole pines were classified as serotinous or nonserotinous. We concluded that the percentage of serotinous trees is most variable at intermediate scales of 1–10 km, and is relatively homogeneous at both fine scales (<1 km) and at very broad scales (tens of kilometers). Percent serotiny was generally more variable and greater at low to middle elevations. Prefire density of serotinous trees was a more important predictor of postfire seedling density than aspect, slope, or soil type. These findings have important implications for landscape-level patterns in postfire regeneration of lodgepole pine.