We now have many connections regarding animal species, such as whales, elephants and bovines. My understanding of these connections is that when visiting a listed WHS, you will be able to see a living creature of that species. Some already have that description (f.e.: "WHS were elephants, Asian or African, live in the wild.").

Lately I have been getting suggestions also to include fossils of those species.
I tend to reject them, as they are not alive anymore (and also often not in situ also, transported to a museum outside the WHS boundaries). Especially Messel Pit was full of these.

On the other hand, it could be worthwhile to group these fossils in some way.

Can someone think of a valuable solution?

Yes, I had never noticed that we don't actually have any "Connections" for fossils other than for "Early Hominid Remains." I notice also that only sites with humanoid fossils have been given a "timeline" period – always "Pleistocene" since we only assigned a timeline to sites which were cultural or mixed - non humanoid fossil sites are categorised as "Natural" by UNESCO.

We do have a "Category" of "Geological Formation – Palaeontological". But not all "Fossil" sites have this category. Ngorongoro for instance is categorised as a "Wildlife Habitat" since Olduvai is only a part (added later) of its OUV

So we have 17 sites in the "Palaeontology" category
Atapuerca
Australian Fossil Mammal Sites (*)
Chengjiang Fossil Site (*)
Dinosaur Provincial Park (*)
Fossil Hominid Sites
Ischigualasto / Talampaya (*)
Joggins Fossil Cliffs (*)
Lower Valley of the Awash
Lower Valley of the Omo
Messel Pit (*)
Miguasha Park (*)
Monte San Giorgio (*)
Mount Carmel Caves
Peking Man Site
Sangiran Early Man Site
Wadi Al-Hitan (*)
Willandra Lakes

And 12 "Connected" to "Early hominid remains"
Atapuerca
Grottoes of Vezere (**)
Fossil hominid sites
Lake Turkana (**)
Lenggong (**)
Vally of Awash
Valley of Omo
Mt Carmel
Ngorongoro (**)
Peking man
Sangiran
Willandra

* = Paleantological site not also "Early Hominid" = 9
** - Early Humanid not also "Palaeontological" = 4
so 21 "Fossil sites" in all.
There could be more sites not identified above which have some "fossilised" aspects. And it is possible (Likely?) that some of the Hominid fossil sites might also contain fossils of some other creatures

The definition of "Fossil" might be worth a quick look – Wiki says there is no official definition but it is usually used to describe stratified remains older than 10000 years. Ie covering a period from c 2.4 BILLION years ago to just 10000!! This covers a wide range of "types" of fossil from the truly "mineralised" through to the merely "covered" and little different from bones which could be excavated in any archaeological dig.

We could simply introduce a connection for "Non Hominid fossils" (excluding fossilised creatures which happened to be found in primarily hominid fossil sites??). Assuming no overlap or as yet unidentified sites this would comprise the 9 identified above. Is it really worth subdividing this further? Possibilities include
a. By geological period
b. By type of creature.

My feeling would be that such a subdivision would not be "useful" – some sites wouldn't make a "threesome". It would be worth however identifying in the text the "geological period(s) represented by the "fossils" found at the site - thus permitting an easy comparison

Regarding the suggestion that sites where e.g a fossilised horse or rhino or ...... has been found should be linked to the Connection for the live animal. I tend to agree with Els that this connection should be limited to sites where these animals live today/recently. One problem is deciding whether a "proto-XXXX" is really an "XXXX". In any case the "interest" in the current definition is the knowledge it provides about the NP etc. The fact that e.g Elephants inhabited the area in the geological past may be of interest but relates to a different dimension?

And another aspect is that most fossils aren't in situ anymore, having been transported to a safer place (e.g. museums around the world). That's just the opposite from being able to see a bear in its natural habitat.

So I agree that it should become a connection of its own. I will also remove any existing links such as Altamira for bovines.

We could also do only animal fossils, and then name the animals? Otherwise we would get all the plants too... (they are everywhere)

Naming the "animals" could be a big job and then not be complete - look at Messel. Best to only mention any "famous" ones from the site?
Stromatolites are neither algae nor plant??
Are we limiting the connection to sites where fossils are part of the OUV? If so then why bother whether they are plants, animals or whatever? Are there no fossilised tree sites on the list??

"Are we limiting the connection to sites where fossils are part of the OUV? If so then why bother whether they are plants, animals or whatever? Are there no fossilised tree sites on the list??"

I would agree with Solivagant. For simplicity's sake, it would be preferable to stick to OUV related fossils...