Print the stream out to a file, and look at it -- it may be that what you see in a web browser is different from what the program sees, due to browser sniffing techniques, content negotiation, etc.

Just because you get an image in your web browser doesn't mean that your program's going to get an image. (they don't look to have a robots.txt, but it's possible they have some other way of slowing down requests hitting their server too quickly)