2004 Field Season: Day 34
My good friend and colleague Alan Deino joined us this weekend and spent some very busy days walking around the basin looking for new tuffs to sample. I already described how we date tuffs using the argon-argon method, and how useful it is for archeologists. I didn't mention, though, that not all tuffs can be dated with this method. In order to get a good date from a tuff, you need to have large crystals of feldspar (a kind of mineral that collects argon). If there are no feldspar crystals in the tuff, or if the crystals have been damaged or contaminated, you can't date the tuff.
What is the use of an undatable tuff? Well, it's not a dead loss - you can still use a method called chemical correlation to match the tuff that's near your site to the same tuff somewhere else, where there are crystals to date.
Basically, whenever a volcano erupts, the ash and pumice that come out all have the same ratio of chemicals. This chemical composition is the signature or fingerprint of every tuff from that eruption, so if you find a tuff in one area with a particular signature and another tuff in another area with the same signature, you can be pretty sure they were from the same eruption. Chemical correlation allows us to compare tuffs across the whole basin, and figure out which layers match up with others, without having to date everything.
Al is our expert not only on argon-argon dating but also on chemical correlation. Here's how it's done. First he needs to find fresh glass in the tuffs. Remember, obsidian is volcanic glass, but pumice is also glass, just with a lot of air bubbles in it. He crushes the sample and treats it with acid to remove any carbon contaminants. He then continues to purify it by using magnetic and density techniques, until the sample is 99.9% pure glass. Glass is not magnetic and it floats, so it's easy to separate.
When it's finally ready, Al ships the sample to a reactor facility in Missouri, where they are able to measure the ratios of different earth elements in the glass, like iron, terbium, zirconium, leutetium and cerium. Each eruption has a unique ratio of these elements, so Al can compare samples from one area of Olorgesailie to other samples he's taken in the past. There are about fifty known eruptions that left chemical evidence in different areas of the Olorgesailie basin. But there are also about 12 to 15 'orphans', or tuffs that don't match to any other ones.
Al looked at the tuffs in Locality G to help Alison and John date one of their sites, GNF-1, where there are many artifacts that look older than the Acheulean, but not quite like the Middle Stone Age. There's a tuff just below this artifact layer, but it's very fine-grained and cannot be dated. Al took a sample for chemical analysis, and he hopes to match it to other samples of tuff elsewhere in the region. Even if he can't match it right away, knowing the chemical signature of this tuff now will help build the library of eruptions here in the Olorgesailie basin, and perhaps later we will find a matching tuff that we can date.