Hi Steven, that is an interesting idea which may provide some valuable insights into refuge use.

One of the problems of using infra-red triggered camera traps for reptiles seems to be that in order to trigger the trap the subject needs to be warmer than the surroundings. For reptiles this might not be the case.

This and other limitations are discussed in 2 papers (linked below but freely found through Google):

The Use of Passive Infrared Camera Trapping Systems in the Study of Frugivorous Monitor Lizards (Bennett & Clements, 2014)

Can Camera Traps Monitor Komodo Dragons a Large Ectothermic Predator? (Ariefiandy et al., 2013).

I also imagine that slow movement and partial concealment by vegetation when entering a refuge might further reduce the success rate. However, there is an interesting article (linked below) where a light beam is used as the trigger.

Hope that helps.

http://www.earthtouchnews.com/discoveries/discoveries/move-over-mammals-threatened-amphibians-need-camera-traps-too

http://varanidae.org/8-1-complete_issue.pdf

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0058800

As Darren says, PIR sensor is dependent on movement + heat difference. I would recommend a standalone mini DVR with video motion detection (VMD) coupled with a small CCTV camera. I have used it for monitoring hawk moths pollinating Orchids, bumblebees visiting flowers and cavity use by lobsters. This system register pixel change in the frame captured by the camera, hence all moving object will trigger the recordings. One more thing is that you could mask area and define the area that triggers the recordings. The size of the study organism is not limiting, you only need to use different lenses on the camera. The smallest object I have video monitored by the use of VMD is small flies and ants. I have attached 3 papers where the use of a portable mini DVR w/VMD is described. 

For a small defined area such as nesting hole, nest boxes, singly plants/flowers and cavities the mini DVR w/VMD is advantageous. And for insects, fish, crustaceans, amphibians and reptiles I would say that this is the best options. 

More details:

The system: The video motion detection (VMD) is in the software of the mini digital video recorder (Mini DVR) which store video on a SD-card when there is movement. The mini DVR is standalone unit. The camera, any kind of analog CCTV camera, is connected to the mini-DVR. The camera is active all the time, but the mini-DVR only record when there is movement.

The VMD: video motion detection is based upon changes in the pixels of the film from the camera. When there is movement the pixels will change, and the magnitude of the pixel change fully depends on the relative size of the moving object. Also shape and colour contrast is important parameters.  When monitoring very small animals a magnification lens is required and only a small surface could be monitor. I do not have a rule of thumb about what the relative size of the moving object should be for optimal monitoring, but just anecdotally I would say that 10% of the monitoring frame.

Regards,

Ronny

http://onlinelibrary.wiley.com/doi/10.1111/j.1756-1051.2012.01574.x/abstract

http://www.pollinationecology.org/index.php?journal=jpe&page=article&op=view&path%5B%5D=114

http://www.publish.csiro.au/paper/MF13139.htm

Hey guys, thanks for the suggestions and recommended reading. When I am free next week I am going to grab a tea (or two) and give them a read. I realise that trying to use PIR to detect reptiles isn't the easiest thing. Hopefully the suggestions you have made will help me find a solution.

Some cam traps (eg Reconyx) have time lapse settings, where a picture can be taken every X hours/minutes no matter what.  Depending on the situation (how long the animal will stay in the refugia, how often you can refresh the memory and batteries, etc.) you might be able to use that.

Congratulations on the idea. There are no studies on this topic with reptiles, and would prove very exciting in natural environments. 

You couldn't assume a passive infra red system would always detect your lizards and snakes. A time lapse system on a very brief interval that excludes the possibility of animals running in and out of the shelter and out of the detection field between shots, or continuous video, would be more reliable, preferably mounted above the shelters if possible.

someone is doing this for she oak skinks in the Snowy mountains - i see the cameras next to roof tiles (commonly used skink refugia) in  our study areas regularly , ask nick Clemman I think he knows

 Hi Steven, have a read of 

Welbourne, D. (2013) A method for surveying diurnal terrestrial reptiles with passive infrared automatically triggered cameras, Herpetological Review, 44(2), pp. 247-250.

My PhD thesis is on developing a camera trapping method for reptiles. I have a couple of other pubs in review/ press at the moment on the method also.