Mumba Cave, located near the highly alkaline Lake Eyasi in Karatu District, Arusha Region, Tanzania. The cave is a rich archaeological site noted for deposits spanning the transition between the Middle Stone Age and Late Stone Age in Eastern Africa. The transitional nature of the site has been attributed to the large presence of its large assemblage of ostrich eggshell beads and more importantly, the abundance of microlith technology. Because these type artifacts were found within the site it has led archaeologists to believe that the site could provide insight into the origins of modern human behavior. The cave was originally tested by Ludwig Kohl-Larsen and his wife Margit in their 1934 to 1936 expedition. They found abundant artifacts, rock art, and burials. However, only brief descriptions of these findings were ever published. That being said, work of the Kohl-Larsens has been seen as very accomplished due to their attention to detail, especially when one considers that neither was versed in proper archaeological techniques at the time of excavation. The site has since been reexamined in an effort to reanalyze and complement the work that has already been done, but the ramifications of improper excavations of the past are still being felt today, specifically in the unreliable collection of C-14 data and confusing stratigraphy.

In 1938 Margit Kohl-Larsen returned to Mumba Cave in an effort to obtain a more complete understanding of the site. Although her work was considered a great triumph in African archaeology at the time, more recent work at the site has illuminated issues from the excavation. Most of the problems lie with the lack of study of the site's uncovered artifacts, soil samples, and faunal remains. Additional problems lie in that very little information from the site, most of which is now considered unclear, ended up being published by Ludwig Kohl-Larsen in 1943.

Further excavations have revealed that there was a significant disregard for the lithic assemblage at the sites, making it clear that Kohl-Larsen favored large lithics above secondary or tertiary flakes. It has been calculated that only the largest 2% of the stones were collected. However, this is unsurprising since the presence of microliths was not yet realized as a hallmark of the Late Stone Age. All problems considered, archaeologists such as M.J. Melhman have stated that the importance of the site should not be downplayed.

During the Kohl-Larsen excavation the stratigraphy and levels of the site were determined within its nine by twelve and a half meters area. Depth of the excavation varies between sources as bedrock has said to have been to be reached between nine and eleven meters. The site was divided into six beds that are now recognized as potentially problematic but still used for reference in present work and reanalysis. In later excavations, artifacts from each of the defined beds have been dated using radiocarbon dating techniques. Unfortunately, it has been difficult to confirm and fully understand the site due to heavily mixed deposits from the Kohl-Larsen excavations, few comparable sites, and lack of data on the past environmental conditions. Each bed has been C-14 dated; however, some scholars question these dates.

Bed I is referred to as the surface level of the site. It has been noted to contains remnants of obsidian, ostrich eggshell beads, fireplace wood, pottery sherds, and arrowheads that are similar to those fashioned by the Hadza people.

Artifacts from Bed II have been dated using radiocarbon to 381 ± 91 B.P. Amongst rockfall debris, this bed contains both human and animals bones, including the bones of turtles and fish, some of which have been carved. Other artifacts include ostrich eggshell beads, pottery, obsidian tools, red ochre.

Bed III contains a particular rich assemblage with the presence of a kitchen midden made of several animal bones and broken pottery. The bed also includes decorated ostrich eggshell beads and pottery. Additionally several human burials found at the site were discovered within this bed. The ceramic artifacts found in this bed are attributed to a type known as "Kansyore Ware", associated with hunter/gatherer peoples. Also found were Neolithic "Narosura Ware" and Iron Age "Lelesu Ware." This bed has an associated radiocarbon date of 844 ± 78 B.P.

This Bed is recognized by archaeologists as sterile containing a large beach deposit. Dates from available material are 21530 ± 320 B.P.