Hobbits, in this very real instance, is the colloquial term for an extinct small bodied ancient relative of the human race Homo floresiensis. Their bones turned up in 2003 at Liang Bua, a limestone cave on an island in eastern Indonesia. Adults stood about a meter tall.
For years the story seemed to stretch almost to the present. Early dates put them in the cave from roughly 95,000 years ago to as recently as 12,000 years ago. That long span made contact with modern humans look likely. Later work on the cave layers shortened the story. The skeletons now sit between about 100,000 and 60,000 years ago. Stone tools and the large animals that lived beside them fade near 50,000 years ago. Modern humans leave a clear trace at Liang Bua around 46,000 years ago.
The gap matters. No one has shown that the two human groups shared this cave. A 2025 climate study now puts a dry turn at the center of the last chapter at the site. Summer rain collapsed around 61,000 years ago. The last dense scatter of dwarf elephant bones and the last hobbit skeletons sit on that same change.
The Hobbits Environment
Homo floresiensis lived on Flores for a very long time. Related fossils from the So’a Basin, about 80 kilometers east of Liang Bua, show that dwarf elephants and Hobbits shared the island for hundreds of thousands of years. At Liang Bua the hobbits left skeletons, stone tools, and a fauna that included Stegodon florensis insularis, a dwarf elephant that weighed several hundred kilograms. Giant storks and vultures appear in the same layers. The elephant was an important prey animal. Most of the Stegodon bones at the cave belong to young animals. Earlier excavation teams argued that the hobbits hunted or scavenged those juveniles. Both species needed surface water. Living elephants drink heavily and suffer when dry seasons drag on. The Hobbits who took their young needed the same rivers and pools.
Why the dates at Liang Bua changed
The first published ages mixed cave deposits that later work pulled apart. Revised stratigraphy and new dates reset the occupation. All Homo floresiensis skeletal remains now fall around 100,000 to 60,000 years ago. The stone tool record and the classic large fauna that travel with the hobbits end near 50,000 years ago. Modern humans appear later. Around 46,000 years ago the raw materials used for tools change. That shift marks a different group at the cave. Definitive overlap between the two species is still missing at Liang Bua and elsewhere on Flores. The new rainfall history sits inside that revised timeline. It covers 91,000 to 47,000 years ago, which includes the last clear phase of hobbit and Stegodon activity at the site.
Why freshwater sits at the center of the problem
Flores has a wet summer and a long dry season. Today about 60 percent of the yearly rain around Liang Bua falls from December through March. The rest arrives with trade winds and storms in the other eight months. Mean annual rainfall near the cave is about 1,400 millimeters. The Wae Racang river drains the valley below the cave. In wet years it can hold base flow for much of the year. In dry years that flow fails. A climate that cut summer rain would shrink the river when animals and people needed it most.
Older climate records from Flores used oxygen isotopes in cave calcite. Those records track monsoon circulation well. They do not cleanly separate how much rain fell from which season delivered it. Summer storm rain and winter trade wind rain carry different isotope values. A change in the mix can move the isotope number even when the yearly total stays the same.
The stalagmite from Liang Luar
The climate record does not come from Liang Bua itself. It also comes from Liang Luar, a cave about 660 meters away in the same limestone hills. A stalagmite called LR09-K2 grew more than 860 meters inside the cave and stands 1.54 meters tall. Thirty-two uranium thorium dates fix its growth from 91,400 to 46,800 years ago. The typical variable on those ages is a few centuries. Magnesium, strontium, and oxygen were measured on the same sample powders, so the chemical records line up in time.
Liang Luar is a good cave for this work. Humidity stays near 100 percent. Evaporation on the stalagmite surface is limited. Other stalagmites from the same cave give similar oxygen values across the last 91,000 years. That agreement means the calcite is recording the amount weighted average of drip water, not a freak local effect in one passage.
What all these tests show is a country getting drier across a long span, then summer rain falling to a low around 61,000 years ago. In the life of a cave occupation that had lasted tens of thousands of years, that is a tight turn.
When the rains left
Modeled mean annual rainfall drops from about 1,560 millimeters in the interval 83,000 to 76,000 years ago to about 990 millimeters in the interval 61,000 to 55,000 years ago. That is a decline of about 37 percent. The early figure sits a little above the modern average. The latter figure sits well below it. Summer totals fall harder than the yearly mean. Early in the record, summer rain reaches about 1,030 millimeters. In the window 61,000 to 55,000 years ago it averages about 450 millimeters. That is about 48 percent below the modern summer amount at the site. Winter rain was high around 80,000 years ago, during a warm interval, and low around 66,000 years ago, during a cold glacial stage. Those winter swings change how seasonal the year feels, but they do not cancel the later crash in summer rain.
The authors compare the wet early climate with rainy country in southeastern Sulawesi. They compare the cold stage near 68,000 to 65,000 years ago with savanna in northern Australia. They compare the driest summers after 61,000 years ago with subtropical uplands on the east coast of Australia, at the outer edge of the Australian summer monsoon.
The fossil record follows suit
Tooth oxygen rises and falls in step with the stalagmite. Low values mark a wet interval near 90,000 years ago. Values climb through the next warm stage. They drop again in the cold stage that follows. A linear fit between excavation depth and the dated cave record places most of those teeth between about 91,000 and 61,000 years ago.
The refined age and depth scale tightens the fossil clock. In two well documented trenches, 92 percent of 716 Stegodon specimens fall between depths that correspond to 76,000 and 62,000 years ago. Only ten specimens are younger than 62,000 years. The youngest dates to about 57,000 years. Homo floresiensis bones are much rarer. They last appear around 61,000 years ago.
Conclusion
The 2025 study fixes a climate date that earlier work at Liang Bua lacked. For fifteen thousand years the country around the cave lost rain. After 61,000 years ago the summer monsoon that now supplies most of the year’s water fell to a level the authors compare with the dry edge of the Australian monsoon.
The last thick scatter of dwarf elephant bones and the last hobbit skeletons sit on that turn. Water is not a side issue for an elephant that drinks like a living elephant, or for a hunter who takes its young. When dry season flow in the Wae Racang failed, the cave that had held both species for tens of thousands of years lost its advantage.
