We have conducted a VLA search for radio pulsars at the positions of 4
4 nearby OB runaway stars. The observations involved both searching im
ages for point sources of continuum emission and a time series analysi
s. Our mean flux sensitivity at 1.4 GHz to pulsars slower than 50 ms w
as 0.2 mJy. No new pulsars were found in the survey. The size of the s
urvey, combined with the high sensitivity of the observations, sets a
significant constraint on the probability, f(p), of a runaway OB star
having an observable pulsar companion. We find f(p) less than or equal
to 6.5% with 95% confidence, if the general pulsar luminosity functio
n is applicable to OB star pulsar companions. If a pulsar beaming frac
tion of 1/3 is assumed, then we estimate that fewer than 20% of runawa
y OB stars have neutron star companions, unless pulsed radio emission
is frequently obscured by the OB stellar wind. Our result is consisten
t with the dynamical (or cluster) ejection model for the formation of
OB runaways. The supernova ejection model is not ruled out, but is con
strained by these observations to allow only a small binary survival f
raction, which may be accommodated if neutron stars acquire significan
t natal kicks. According to Leonard, Hills and Dewey (1994), a 20% sur
vival fraction corresponds to a 3-d kick velocity of 420 km s(-1). Thi
s limit supports recent revisions of the pulsar velocity distribution.
(C) 1996 American Astronomical Society.