Okay. I saved those same two pngs as bmps; deleted the 3 old zips; and then compressed the bmps per your post:

C:\test>dir ?.bmp
31/05/2011  14:33           720,054 1.bmp
31/05/2011  14:33           720,054 2.bmp
               2 File(s)      1,440,108 bytes

C:\test>del 1.zip 2.zip 3.zip

C:\test>zip 1.zip 1.bmp
  adding: 1.bmp (160 bytes security) (deflated 100%)

C:\test>zip 2.zip 2.bmp
  adding: 2.bmp (160 bytes security) (deflated 100%)

C:\test>zip 3.zip 1.bmp 2.bmp
  adding: 1.bmp (160 bytes security) (deflated 100%)
  adding: 2.bmp (160 bytes security) (deflated 100%)

C:\test>dir ?.zip
31/05/2011  14:34             2,376 1.zip
31/05/2011  14:35             2,376 2.zip
31/05/2011  14:35             4,730 3.zip
[download]

(2376 + 2376) / 4370 = 1.0874141876430205949656750572082.

Still nowhere near the 2.0 you would expect if this worked.

Concatenating the bmps doesn't help much either. I got 1.12

Perhaps we need a theoretical perfect compression algorithm?

Great minds :)

C:\test>copy 1.bmp/b+2.bmp/b 3.bmp
1.bmp
2.bmp
        1 file(s) copied.

C:\test>del 3.zip

C:\test>zip 3.zip 3.bmp
  adding: 3.bmp (160 bytes security) (deflated 100%)

C:\test>dir ?.zip

31/05/2011  14:34             2,376 1.zip
31/05/2011  14:35             2,376 2.zip
31/05/2011  14:47             4,472 3.zip
[download]

---

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.