The size distribution of short-period comets, although crucial
 for assessing theories of their origin (e.g. whether from the
 Kuiper belt, the inner core of the Oort cloud or the observed
 long-period flux), remains largely uncertain.  Most current
 estimates of cometary diameters depend on assumed visual
 albedos together with nuclear magnitude estimates, or on
 arguments based on outgassing due to water-ice sublimation and
 assumptions about the fractional active surface area. Accurate
 measurements of the size depend on infrared measurements in
 the 10 to 100 micron range (depending on the temperature of
 the nucleus), and currently only some 5 objects have reliably
 determined diameters. The sensitivity of ISO provides an
 opportunity to observe a substantial number of Jupiter-family
 comets, currently moving in a variety of typical short-period
 cometary orbits. We propose to use ISO to observe a number of
 these comets relatively far from the sun in order to determine
 the sizes of their nuclei while they are comparatively
 inactive and uncontaminated by coma. The proposed program
 will more than double the sample of accurate diameters of
 short-period cometary nuclei, and combined with contemporary
 visual observations will also allow accurate determinations
 of their visual geometric albedos.