Several techniques for diagnosing adenocarcinoma of the lung are commonly available, but the frequency of their use and diagnostic sensitivity may vary. Twenty cases of primary lung adenocarcinoma obtained at surgery were studied by the following four routine techniques: light microscopy (LM) using hematoxylin-eosin (H&E) stain, mucicarmine stain, and PAS-diastase stain, and electron microscopy (EM). Three observers independently determined the positivity (0 [none], +/- [equivocal], 1 + [slight], 2 + [moderate], 3 + [marked]) of each of these cases for lumen formation in H&E-stained sections (LM lumens), intracytoplasmic (cytoplasmic mucicarmine) or intraluminal (luminal mucicarmine) mucicarmine, intracytoplasmic (cytoplasmic PAS) or intraluminal (luminal PAS) PAS-diastase, and lumen formation (EM lumens) or microvilli (EM microvilli) on electron microscopy. Comparative matching of these seven microscopic determinants (using Wilcoxon signed-rank test) demonstrated significant (P less than .01) sensitivity of EM microvilli over EM lumens, EM microvilli over luminal mucicarmine, cytoplasmic PAS over luminal mucicarmine, EM microvilli over cytoplasmic mucicarmine, cytoplasmic PAS over cytoplasmic mucicarmine, and EM microvilli over LM lumens, and a significant (P less than .05) sensitivity of cytoplasmic PAS over LM lumens, EM microvilli over luminal PAS, luminal PAS over luminal mucicarmine, and cytoplasmic PAS over EM lumens. Friedman's nonparametric test (P less than .05) indicated a significant difference among the microscopic determinants. The most sensitive was EM microvilli (mean rank score, 5.17) followed by cytoplasmic PAS (4.77), luminal PAS (4.02), cytoplasmic mucicarmine (3.62), LM lumens (3.52), EM lumens (3.47), and luminal mucicarmine (3.40). However, each of the diagnostic techniques had case examples positive for one, but not for the others, indicating that maximum yield of adenocarcinoma diagnoses will be obtained by performing all four techniques (H&E, mucicarmine, PAS-diastase, and electron microscopy.