Background Two contradictory hypotheses have been proposed to explain the relationship between allergic conditions and malignancies, the immune surveillance hypothesis and the antigenic stimulation hypothesis. The former advocates that allergic conditions may be protective against development of cancer, whereas the latter proposes an increased risk. This relationship has been studied in several case-control studies, but only in a few cohort studies. Methods The association between allergic conditions and risk of developing leukemia, Hodgkin's disease, non-Hodgkin's lymphoma and myeloma was investigated in a cohort of 16,539 Swedish twins born 1886–1925. Prospectively collected, self-reported information about allergic conditions such as asthma, hay fever or eczema was obtained through questionnaires administered in 1967. The cohort was followed 1969–99 and cancer incidence was ascertained from the Swedish Cancer Registry. Results Hives and asthma tended to increase the risk of leukemia (relative risk [RR] = 2.1, 95% Confidence Interval [CI] 1.0–4.5 and RR = 1.6, 95% CI 0.8–3.5, respectively). There was also an indication of an increased risk of non-Hodgkin's lymphoma associated with eczema during childhood (RR = 2.3, 95% CI 1.0–5.3). Conclusion In contrast to most previous studies, our results do not indicate a protective effect of allergic conditions on the risk of developing hematological malignancies. Rather, they suggest that allergic conditions might increase the risk of some hematological malignancies.