First: please, keep in mind the fact that the [i]estimates[/i] which I use below, are very, very much skewed in favor of the "all are rocks" theory.
I use two convenient objects:
(a) manta ray with tail
(b) fish statue
I will prove that probability that they are "simply rocks" is less than 1 in 1,000,000,000,000,000 !
Lets do some math, using the "manta ray with tail" picture.
Lets assume that there is 10,000,000,000=10[up]10[/up] rocks that can be considered for fossils in the area covered by rovers
Lets assume that each of them has a small rock attached, which will play the beginning of the tail
Lets assume that there is 1/100 chance that small rock has an another small rock attached on the right side of it and having right size, so that it forms next bone of the tail. A so on, and so forth.
Since the "tail" has 12 bones, expected number of existing in the area "rock stingrays" will be 10[up]-14[/up]
Lets assume that rovers photograph everyone of them, that exists.
We must conclude based on the assumptions and their consequences, that chances of us to have the existing picture if the object were of nonbiologial origin, would be less than 10[up]-14[/up], which is hundred thousand times less than 1 in 1,000,000,000
(b) Lets assume, that probability of photographing of fish like rock on mars is 1
and that there is 1,000 of them in the rovers covered area
Now, let us focus on the string of similar dots leading to the fish tail. There is 7 of them.
Lets assume that chances that such line on the simple rock has one extra dot are 1 in 100. After all, I've never seen even three dots like this after viewing probably 10,000
rocks /which corresponds to 2 additions/.
The chances we would have a photo of the fish like rock would be than 10[up]-9[/up]
which is 1 in 1,000,000,000 as expected.
Whichever way you slice it, chances that we would have pictures we now have if "they all are rocks" theory were correct, are negligible.