We investigated behavioral symptoms of neuropathic pain, and associated changes in dorsal horn neurons, in a rat model involving loose ligation of lumbar dorsal roots. The L4–L6 dorsal roots were exposed unilaterally and loosely constricted central to the respective ganglia with one (1-ligation) or two (2-ligation) silk 7-O ligatures. In control groups the dorsal roots were exposed but not ligated (sham-operated), or sutures were placed lengthwise between the dorsal roots (suture control). There was a significant reduction in mechanical withdrawal threshold on the operated side in both 1- and 2-ligation groups which began at 3 days, peaked at 2–5 week, and gradually recovered. A delayed threshold reduction was also seen on the non-operated side. Immediately post-surgery there was a significant increase (hypoalgesia) in thermal paw withdrawal latency (Hargreaves test) in 1- and 2-ligation groups on the operated (but not non-operated) side that recovered after 1 week. Significantly less weight was borne by the operated limb 1–5 weeks post-operatively in 1- and 2-ligation groups. The force of hind limb withdrawals elicited by graded noxious heat pulses (38–52°C) was significantly lower 1 week post-surgery on the operated side (1-ligation group) followed by recovery. Withdrawal forces were higher 5–9 week post-surgery on the non-operated side in 1- and 2-ligation groups. We found no evidence of cold allodynia. Neither sham-operated nor suture controls showed any signs of allodynia or hyperalgesia. Following behavioral testing, rats were anesthetized with halothane for single-unit recordings from lumbar wide dynamic range-type (WDR) neurons. At 22 week post-surgery, the mean area of mechanosensitive receptive fields was significantly larger for units on the operated side in 1- and 2-ligation groups compared with those on the non-operated side or with those from sham-operated rats. Mean stimulus-response functions to graded noxious heat pulses (38–52°C, 5 s) were not significantly different between operated and non-operated sides for 1- or 2-ligation groups, or compared with the 22-week sham-operated group. At 5 week post-surgery, the mean area of cutaneous receptive fields, and stimulus-response functions to graded noxious heat, were not significantly different between units recorded on operated versus non-operated sides, or compared with units from 5-week sham-operated rats. Spontaneous unit activity was significantly higher on the operated versus non-operated side in the 2-ligation (22-week) and sham (5-week) groups. Enlarged cutaneous receptive fields of dorsal horn neurons may contribute to mechanical allodynia associated with dorsal root constriction. However, the slow (>5 week) development of receptive field enlargement does not match the rapid development of allodynia. The lack of effect of dorsal root constriction on thermal sensitivity of dorsal horn units ipsilaterally corresponds to the lack of marked thermal hyperalgesia observed behaviorally.