This is an existing connection http://www.worldheritagesite.org/tag.php?id=205.
I thought it be worthwhile to make it more exhaustive, however, during my work on it I came to a conclusion there are some details in need of discussion before doing so.

1) Eurocentricity: What do we exactly mean by discovered? Angel Falls for instance were well known to the indigenous population of Venezuella before they were "dicovered" by Angel. The same is true of many of the listed sites (Tsodilo, Uluru, Grand Canyon, Kilimanjaro...). Do we mean first visited by Europeans? This should then be a distinct connection, I think.
Interestingly, the reverse seems absurd. Rabban Bar Sauma was the first East Asian to visit Europe. http://www.worldheritagesite.org/tag.php?id=696
We could then readily argue that for the Chinese he discovered Paris, Rome, Naples and Istanbul!

2) Archaeological sites: Very few archaeological sites remained visible and known in their surroundings (e.g. Acropolis, Borubadur, Angkor, Pyramids, Roman Forum). All the rest, which are certainly more than 200 sites in the list had to be excavated. Do we consider archaeological excavation a discovery?

3) We already have a connection Recently discovered http://www.worldheritagesite.org/tag.php?id=411.
In order to avoid an overlap we need to define years for Recorded discoveries too.

4) What do we do with Secret locations that got exposed? http://www.worldheritagesite.org/tag.php?id=448
A good example is Auschwitz which was "discovered" by the Allies at the end of WWII. Unfortunately, it was well known to others, including many Europeans.

What do you think?

A few observations if not total "conclusions"!
a. What do we exactly mean by "discovered"?
"Not previously known to exist either at all or by persons beyond the local population" – this normally relates to a geographical (Island/mountain etc) or historical place but could refer to the significance of an already "known" location where its significant aspect was not previously known (e.g An archaeological site whose significance is only established by excavation and study). So Borobodur was genuinely "discovered" in its entirety when knowledge of its existence was made available beyond the locals through the survey ordered by Raffles. On the other hand some ruins at Moenjodaro were known to exist but without any knowledge or understanding of what they were until the existence of a significant Indus valley city was "discovered"

b Do we mean first visited by Europeans?
"Visiting" isn't the same as "discovery". Places can be "known" to exist by persons beyond the locals even though they have been visited by very few or the knowledge of them is very limited (e.g because it is only recorded by some secret map). I don't think the potential example given of Rabban Bar Sauma "discovering" Paris etc as far as East Asians are concerned is relevant to the Connection as currently being defined since no one has suggested that e.g Marco Polo should be recorded as having "discovered" places in Cathay just because he "visited" and was the first European to write about them. An example might be that of Galle – its existence was well known by many civilisations continuously through "historical times" – I don't think it would be fair to classify as a "discovery" either the visit by Zheng He in 1411 or by Lorenco de Almeida in 1505 (the first recorded Chinese and European visitors) – and we haven't. The historical reality is however that, because of the nature of World exploration, the first recording of places previously unknown at all or only known to locals is most likely to have been done by "Europeans".

c. Do we consider archaeological excavation a discovery?
If the excavation is the first on the site and thereby establishes the nature of the site, and what was to become its OUV etc then surely that would represent a "discovery" in that it wasn't previously "known" for what it was. Later excavations which may well discover further valuable artefacts or information about a site would not count. Moenjodaro seems a useful example since it also involves a "discovery" by a non European – the existence of "something" on the site was well known but the site's "emergence" as the remains of an entire Indus Valley city which resulted in it being given its name ("Mound of the Dead") can be assigned to Bandyopadhay in 1922. We certainly could Connect Archaeological site discoveries separately from Geographical site discoveries. That could reduce logical problems identified with other Connection types as below and also improve clarity by splitting what could be regarded as different "populations". The Archaeological discovery connection would encompass either the first identification of the site altogether OR its first identification as somewhaere significant – whichever was relevant. The same rules would apply to knowledge by locals as for Geographical discoveries

d. What of Recent Discoveries?
Nearly all of these sites are Archaeological ones, though the inclusion of the Monarch Butterfly Reserve and additional lava tubes at Jeju relate to "scientific" discoveries. There does seem some benefit in terms of specific interest in identifying "recent" discoveries but I agree that many of them are not much different from the general "archaeological" discoveries and the date can be provided to highlight the "recent" ones (a subjective term anyway). On balance I might put the discovery of where Butterflies went in the winter to "Scientific developments" and the Jeju tubes to Geographical discoveries. By the way Els - the Cuava de los Manos discovery date isn't as per the Nom file – so I don't know where it came from

f "Secret Locations".
The "secrecy" aspect is surely different from that of "discovery". Most "discovered" places have not deliberately been kept "secret". Some sites might of course fit both Connections if the secrecy was so great that their existence was unknown other than to locals. Machu Picchu could be such a site (if we accept that its location was indeed kept deliberately "secret" rather than just being "lost" during the disintegration of the Inca Empire. In reality I don't think we know this but it is all part of the "MP myth"!). I think that a number of other entries on this list also shouldn't really be there. "Hiding places" should perhaps be given their own Connection (including the Anne Frank house) - this then clearly differentiates them from the "Discovery" aspect. Why does the fact that some caves on Rapa Nui were hidden by walls make them a "secret location" – we don't really know that.

I always thought that the focus of this connection was on RECORDED discoveries. So: an individual discovers a site, and writes about it himself (or another writer has written about the discovery, and attributes it to the indivual that found it).

Agreed - but that still begs the question as to what is meant by "Discovering"!

Then would Auschwitz fit the definition of Discovery here? It was not widely known before the Allies liberated it.
Otherwise I think the definition brought forwards by Solivagant can serve as an objective criterion that solves the problem of ethnocentricity.
We still need to decide how to exactly split this connection which has the potential of becoming a huge one. I think that Solivagant's idea to split geographic and archaeological discoveries is a good one. Recent discoveries could be maintained separately, whereby each of them would also belong to one of these additional two connections.
As to Els' remark, of course only recorded discoveries would be in.
Any thoughts?

I wouldn't have thought so it WAS known if not widely - but "discovery" can surely only happen if "nobody" (other than the exception of "indigenous people") knows

If we indeed decide all archaeological discoveries should be in we should add quite a lot of sites to this connection. It currently includes: Angkor, Borobudur, Vezere, Hattusha, Machu Picchu, Moenjodaro, Spiennes, Petra, Nemrut Dag, Troy and Tsodilo. Both Thracian graves are in but without any detail about their discovery.
There are surely more than a 100 WHS that could be added here. So if we are to complete this connection I would be happy to have some help. Could we divide the workload according to geographical areas?
By the way, some archaeological discoveries should be better classified as geographical ones: for example, Easter Island or Tenochtitlan.

So:
- Not previously known to exist either at all or by persons beyond the local population
- Individual(s) that discovered it must be named
- Year of discovery must be named

Geographical discoveries would be:
Bikini, Atlantic Islands, Canaima, Gough, Grand Canyon, Henderson, Iguazu x2, Kilimanjaro, Macquarie, Mammoth, Papahanaumokuakea, Mt Kenya, Phoenix, Easter, Redwood, Shark Bay, Arena Sardona, Uluru, Victoria, Redwood, Yellowstone.

Chengjiang like other fossil sites should be added to archaeological discoveries as no new geographical features were discovered.

From Recently Discovered we could add: Jeju, Monarch Butterfly (whereby I am not that sure as to the latter)

This is in general not what I want with the connections. Originally, we had 3 to 20 as a starting point. Some might be longer, no problem. But this will not result in a real Connection (the 100 individual sites aren't connected at all), it results in a list of features of the sites. Similar to a list of names of architects that constructed buildings that are WHS.

The definition is too broad.

I agree. Most archaeological sites were discovered in recent times and these discoveries were recorded. There is also a huge number of archaeological sites on the list.
If we are to avoid such huge connections I would suggest to remain with two connections only:
Recent discoveries (as we already have)
Geographical discoveries (whereby we also include recent geographical discoveries)
And then omit the archaeological discoveries which are not recent.

I think that could work.
But what is the definition of a geographical discovery? I can understand it includes remote islands. But what about Redwood for example?

What about the Aztec capital Tenochtitlan? Was it discovered by Hernan Cortes? It was not only known to a local population in a small scale. It was the capital of an Empire the influence of which reached as far south as Nicaragua (and maybe even Panama). It was just unknown to anyone outside of Mesoamerica. Does this meet the definition of discovery? If it does, this would be a geographical discovery.

Maybe the connection of Geographical discoveries should include everything that was always exposed (so which did not involve excavation) and is of large scale (so Alta for example would be out), the discovery of a new location so to say.
This would mean Angkor, Monarch Butterfly, Borobudur, Petra and Machu Picchu would still be in as they are places that were unknown before. The same can be said about caves (so Vezere would be in). Moenjodaro and Nemrut Dag would be out as their location was known previous to the archaeological discovery.

mmm, still not convinced, it is difficult to separate the ones that are included and ones that aren't