The density of Halley's comet is determined as a function of a parameterized time of visibility of ejected material, as a function of the angular deviation α from the Sun-comet direction of the averaged reaction force due to jets of ejected material when α is assumed independent of orbital position, and as a function of thermal inertia for a variable α. The variation of the averaged gas ejection rate from the nucleus as a function of orbital true anomaly is estimated by using the lightcurve of D.W.E. Green and C.S. Morris (1987, Astron. Astrophys. 187, 560-568) calibrated by the ejection rate at 1.28 rad past perihelion (0.89 AU) found from the spacecraft flybys. Estimates of the H 2O production rate deduced from several observational data sets are scattered around the gas production rate determined from the lightcurve except very near perihelion where the observationally based estimates are scarce and generally fall below the rate determined from the lightcurve. Although some of the densities obtained by this procedure are in reasonable agreement with intuitive expectations of densities near 1 g/cm 3, the uncertainties in several parameters and assumptions expand the error bars so far as to make the constraints on the density uniformative. So many of these uncertainties apply to all procedures proposed for determining the density that we conclude that secure, informative constraints on the density of any cometary nucleus must await a spacecraft rendezvous with that nucleus. In particular, the suggestion that cometary nuclei tend to by very fluffy, undersense objects should not yet be adopted as a paradigm of cometary physics.