Humans arrived to Cyprus thousands of years earlier than previously thought

A fascinating new study by Bradshaw et al. [1] has provided compelling evidence that the Mediterranean island of Cyprus was settled by pre-agricultural hunter-gatherer populations much earlier than previously thought. The research, recently published in the PNAS, challenges long-held assumptions about the timing and nature of human dispersal to the Mediterranean islands during the Late Paleolithic.

Cyprus is located at the cross-roads of many critical trade routes across the Levant and Near East [2], previous research such as that discussed in our previous article has tried to shed light on the possible routes and journeys these paleolithic sea farers may have undertaken.

The prevailing view has been that the Mediterranean islands were largely uninhabited until the development of agriculture on the mainland, as they were considered inhospitable and unproductive environments for hunter-gatherer societies. However, by combining modern modeling techniques with the latest archaeological data, the authors demonstrate that Cyprus was possibly settled by sizable groups of hunter-gatherers as early as 14,257 to 13,182 years ago. This would be thousands of years prior to current evidence of human settlement on the island such as that discussed in our discussion on Akrotiri.

Modelling the First Arrivals to Cyprus

The simplest way of definitely proving presence of humans on Cyprus is by discovering material that can be accurately scientifically dated, through techniques such as radiocarbon dating. For Cyprus, this would put the oldest date to 11,720 ± 240 years before present. The Signor–Lipps Effect, which relates to how material records being incomplete makes it impossible to accurately determine start and end dates of population events, is a major factor in the uncertainty of the exact date of human arrival to Cyprus.

The CRIWM Method

The authors use a method called calibrated resampled inverse-weighted McInerney (CRIWM), first developed by Bradshaw et al. [3] in 2012 which was designed to estimate the extinction time of a species from well dated records. CRIWM is a statistical technique that analyzes the pattern and frequency of dated archaeological or fossil records to infer when a species or population likely went extinct or first appeared in an area.

The method starts with the McInerny method (2006) [4] which estimates extinction time based on the sighting rate of dated records. This assumes records are equally likely to occur, so a decline in record frequency towards the present may indicate an extinction. In other words, if we see many dated records of a species in the distant past but fewer and fewer records as we get closer to the present, that could be a sign the species went extinct.

The authors modify the McInerny method [4] by inversely weighting each record's contribution to the estimated extinction date based on its temporal distance from the most recent record. This places more importance on the most recent records instead of the oldest and assumes that time between samples is a good proxy for population size. Essentially, this refinement to the method gives more weight to the most recent records, since these are most informative about when the species finally disappeared.

To account for uncertainty in radiocarbon dating of records, the method resamples each record's date by using the samples mean and known error margins. The resampling is repeated many times (e.g. 10,000 iterations) to generate a confidence interval around the final estimated extinction date. Radiocarbon dating always has some uncertainty, so the CRIWM method addresses this by running the calculation many times with slightly different dates within the possible range for each sample. This produces a range of plausible extinction dates rather than a single definitive date.

In the context of the paper being discussed, CRIWM was applied on a dataset of 10 dates. These dates were the oldest known record (usually a radiocarbon dated charcoal sample) from 10 unique archeological sites around the island. Running this with the default settings on CRIWM resulted a median of 13,613 years ago with a 95% confidence interval of 14,257 to 13,182 years ago. This means the analysis estimated humans likely arrived on Cyprus around 13,613 years ago, possibly as early as 14,257 years ago or as late as 13,182 years ago.

While we are not archeologists, we decided to re-run the CRIWM algorithm on the full set of radiocarbon dated charcoal samples available on Cyprus, which amounts to 1022 records. Interestingly, this resulted in an even older age of 13,978 with a 95% confidence interval of 13,642 to 14,787 years ago. Using a larger dataset pushed the estimated arrival date back even further.

While these estimates are wide, they are consistent with the idea that Cyprus was settled by hunter-gatherers much earlier than the oldest well-dated material we currently have. So who were these groups of people and how did they manage to survive on the island?

Minimum Viable Founding Size

The concept of Minimum Viable Population is the approximate minimum number of individuals required for a population to have a high probability of persisting over the long term.

So what was this number? Recall that the oldest boats we know of are small 5 to 10 meter long vessels carved out of single tree trunks. These could hold maybe a few passengers and some cargo. Moving large amounts of humans to Cyprus, in what was likely a multi-day hazardous journey would have been a significant challenge.

At first it's easy to think it might have been only a few dozen people. However this would not result in a stable population. The authors of the paper argue that around 2,750 people would have been needed to have a 90% chance of not becoming quasiextinct. Note that quasiextinction refers to a type of extinction where a population is reduced to a very small size with no possible way of recovering.

The authors further suggest that this founding population likely arrived in 2 to 3 main events, each involving around 1,000 to 1,375 people, occurring within a period of 6 to 98 years. This scenario is more plausible than a single, instantaneous arrival of the entire founding population.

Implications and Future Research

The findings of this study have implications for our understanding of the early human settlement of the Mediterranean islands. They challenge the long-held belief that these islands were inhospitable to pre-agricultural societies and suggest that the maritime capabilities of these ancient peoples were more advanced than previously thought.

However, many questions remain unanswered around the potential for using such methods to estimate population arrivals in varying contexts. Hopefully further archeological research will yield more well-dated sites and artifacts from this period, providing valuable insights into the lives of these early Cypriots.

The study by Bradshaw et al. [1] provides an interesting approach to estimating the early settlement of Cyprus by pre-agricultural hunter-gatherers, challenging traditional views of the island's history. This research opens up new avenues for exploration and highlights the importance of interdisciplinary approaches in the study of ancient human migrations.

We would also like to thank the team for making their code, data and methods freely available on GitHub to allow others to validate and build upon their work. It is essential to the scientific process that research is transparent and reproducible, and we cannot commend the authors enough for their commitment to open science.