I want to know the first point of time in the past when genomes have diverged so far from modern human genomes so that they both (human and ancestors) are not interbreeadable anymore and can be seen as separate species.
Considering only post-zygotic isolation and based on below discussion, I would tend to think that 20th century human could interbreed with our ancestors living a 1 to 2 millions years ago. I would give a range between 100k years and 8 MYA for safety.
Concept of species
The post How could humans have interbred with Neanderthals if we're a different species? makes a good introduction to the difficulties behind the concept of species.
Anatomically speaking, it is probably possible to get a spermatozoid cell from a 20th century man start fusing with a Homo abilis ovule (and vice-versa). The potential pre-zygotic isolation will rather be caused by aesthetic appearance.
To the risk of sounding posh, I would say that Homo abilis hairy face does not exactly fall within my standards! I suspect it will be the same for many of us, humans of the 20th century. It is hard to really know how much time one needs to go back to consider that we don't disgust each other enough to refuse mating together. There are a few reasons why this is hard
- Reconstruction (esp. facial reconstruction) is often too inaccurate to really allow to tell how ugly their faces were.
- It is going to be super hard to tell what they would just about our faces today! I bet they'd laugh at our ridiculously small jaws!
- It is hard to fully separate the environment from the person. In middle-age (in Europe at least) people would bath a few times in a life-time. Again I would consider this not in my standards!
But let's focus on post-zygotic isolation.
The humanzee (Homo sapiens sapiens × Pan) (also known as the chuman or manpanzee) is a hypothetical chimpanzee/human hybrid. An unsuccessful attempt to breed such a hybrid was made by Ilya Ivanovich Ivanov in the 1920s. There have been occasional reports of human-chimpanzee hybridization, notably regarding a performing chimp named Oliver during the 1970s, but none of them have been confirmed. Similarly, the possibility of a chimpanzee–gorilla hybrid, known as a koolakamba, also remains unsubstantiated.
There is no evidence of humanzee but it is important to realize that the idea of such hybrid is not fully unthinkable. Human and chimpanzee share a common ancestor about 6 millions years. So, if a humanzee is consider eventually possible, I would quite definitely consider that a H. abilis (~1 MYA) or H.erectus (~2 MYA) and 20th century human could well interbreed if we were to live at the same time.
Sumatrans and bornean orangutans can hybridize (Malinson, 1984) despite that their lineages have split about 1.3 MYA!
So putting all of that together and follwing some vague intuition, I would personally tend to think that 20th century human could interbreed with our ancestors living a 1 to 2 millions years ago. I would give a range between 100k years and 8 MYA (MRCA with gorillas). But really, it is a rather vague intuition.
See also CrashCourse > How to Make A Humanzee