Thanks for the quick answers .. unfortunately now raises several new questions:
And if you're running delay, you've got the processor running full time too, got the power LED on the whole time, got the at8mu2 running full time, if your unused pins are left floating as inputs you've got potentlal switching losses there (best to set unused pins as inputs with internal pullups enabled).
You'd be way better off running of 3 AAs direct to the 5V header pin, bypassing the regulator.
I am still learning but is there a way to shut down the at8mu2, power LED, and processor for a specified length of time. I can set the unused pins as you suggest but have not done it in this case as all the examples leave all the unused pins alone in the sketches and didn't know this would be a problem.
Last part of your reply was to apply 3 AAs to the 5V header .. 3X1.5(or as pointed out for rechargables1.25)=4.5V(3.75v) .. I was under the impression that you needed to have a 5v supply to the 328p but after looking it up Vcc can be 1.8-5.5v .. does lower voltages effect the ocillator timers and such or the reliability of the chip to operate?
A rechargeable 9V battery reduces the cost over alkaline---but they are normally 150mAh or so, whereas AA's are 2000mAh (13 times the life). And if using rechargeable AA's then use 4, not 3, as they are about 1.25 to 1.3V each rather than 1.5V for alkaline.
Untill you told me I had no idea that there was that much of a difference in mAh between the two types of batteries. For now I am working on a computer upstairs and the breadboard in the basement and thought being able to run the UNO from a "portable 9V supply" would be a good idea. I will just rip a PC supply out of a junk PC and use for 5V supply on the workbench (come to think of it I will have 12V and 3.3V also) Eventually I will have a PC downstairs running Linux with Arduino IDE running on it so will have the USB port downstairs too.
Thanks again guys,