Recent experiments show that the adsorption of biomolecules on one surface of a microcantilever generates surface stresses that cause the cantilever to deflect. If a second species binds to the adsorbed molecules, the stresses change, resulting in a different deflection. By choosing adsorbed probe molecules that recognize specific molecules, it may be possible to detect pathogens and biohazards. In particular, Fritz et al. (Fritz, J.; Baller, M. K.; Lang, H. P.; Rothuizen, H.; Vettiger, P.; Meyer, E.; Guntherodt, H.-J.; Gerber, Ch.; Gimzewski, J. K. Science 2000, 288, 316) and Wu et al. (Wu, G.; Haifeng, J.; Hansen, K.; Thundat, T.; Datar, R.; Cote, R.; Hagan, M. F.; Chakraborty, A. K.; Majumdar, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1560) show that the presence of an individual sequence of DNA may be identified by observing the change in deflection as hybridization occurs. Also, it has been shown that this platform can detect prostate specific antigen (PSA). However, to exploit this phenomenon for the development of reliable microdevices, it is necessary to understand the origin of the nanomechanical forces that lead to cantilever deflection upon molecular recognition, as well as the dependence of such deflections on the identity and concentration of the target molecule. In this paper, we present a model with which we examine cantilever deflections resulting from adsorption and subsequent hybridization of DNA molecules. Using an empirical potential, we predict deflections upon hybridization that are consistent with experimental results. We find that the dominant contribution to these deflections arises from hydration forces, not conformational entropy or electrostatics. Cantilever deflections upon adsorption of single stranded DNA are smaller that those predicted after hybridization for reasonable interaction strengths. This is consistent with results in Fritz et al., but not those in Wit et al. The deflections predicted for DNA before and after hybridization are strongly dependent on surface coverage, as well as the degree of disorder on the surface. We argue that self-assembly of probe molecules on the cantilever surface must be carefully controlled and characterized for the realization of microdevices for pathogen detection that rely on nanomechanical forces generated by molecular recognition.