MIKE SMITH: We’ve worked our way through sediments, layers and interfaces. Now I want to look at features. Features are very neglected in Australian archaeology. I think basically because we dig such small holes and we very rarely can see a feature fully.
When we move to look at features, we’re moving from the sort of sedimentary system that the British would call the natural to a stratigraphy that derives from human construction of the landscape, the sort of material that makes up the bulk of archaeological layers in European sites - construction debris pits, embankments and for which the Harris matrix was really designed for. But here we’re dealing with subtle features. And by features we’re looking at pits, hearths, burrows, post-holes, ash lenses. We’re looking at footprints, animal trackways and so forth.
When you’re looking at a feature, it’s important to look at features from two perspectives. Features are a disturbance, and from one point of view we need to isolate them. If someone has dug a pit in a site, not only is there a possibility of introducing material from a higher level into the deposit but also the fill of that pit has introduced older material and spread it around a younger surface. If you’re in an area where you’ve got a lot of pits, you’re going to have to be very careful about your selection of charcoal for radiocarbon dating.
But a pit is also an opportunity to look at an actual event, not something that took 1000 years to form but an event that might be a moment in time - an hour, a day. It’s worth looking at features in some detail because they may preserve a range of information.
A fireplace might have been cleaned out and reused. The charcoal might give you an indication of the sorts of fuel that was being used, whether people were using acacias or desert oak or they are using something else. A pit that’s been left open on the site might take a little while to fill, and it might accumulate large debris in the same way that secondary disposal areas accumulate. You might find large lumps of ochre. You might find your big cores because they’ve gradually been swept into these depressions in the site.
I think if you’re lucky enough to find a feature, you need to try and expose it not just vertically but also get a sense of what it looks like in plan. You really can’t tell whether that charcoal smudge is half a fireplace just by looking at it vertically; you actually also need to see it in plan. You need get a good sense of its structure. Some fireplaces with internal differentiation of ash lenses and yellow bird sediments and stuff like that are unequivocally fireplaces, but other charcoal smudges might be tree roots. You need to get a bit of a sense of their structure in order to interpret them.
With pits, we need to try to work out what pits have been dug for. At Puritjarra we’ve had little pits that seem to have been dug as fire pits with a fast-burning fuel that has just sooted the edge of a pit. The sort of pit you might dig as part of a process of maybe heating a spear shaft to straighten it or something like that. We’ve had pits that have been infilled with other sediment and lined with rocks. Basically the sort of pits you might build to hold a post - not quite post holes because the post hasn’t been left there. What’s that post for - holding up a structure for women to hold onto when they are winnowing grain?
The features might be animal burrows, and clearly if you’ve got an animal burrow you need to isolate that and keep that separate from the rest of the site, because it will introduce intrusive material.
A feature might be a grave pit. It is important with most features to try to establish from what level they were originally dug. You might be able to determine that. In some cases, if they are cross-cut by a disconformity, the feature itself will be truncated and you won’t be able to work out from what level it was dug. That bit of the section has been lost.
Features: what sort of internal stratigraphy do they have? Can we see them in plan? What’s the surface of origin? What’s the overall spatial patterning of features? We might find sleeping fires towards the rear of a site and the cooking fires towards the front of a site.
Desert archaeologists that have done a pretty good job of planning out features. Dick Gould did a good job of looking at hearths and the spatial distribution of hearths at Puntutjarpa. I’ve done something similar at Puritjarra. But generally we regard features as just disturbance rather than as opportunities.
To give you an example of the biography you can get from a feature, at the rear of Puritjarra there was a little trench dug. It was dug in such a way that it must have been dug fairly quickly, and the soil was pushed back behind the digger back into the pit as the trench was dug forward. I say that because. as it was open, charcoal fell in and lined the walls of the feature and then was quickly buried by other sediment, just giving us a little outline of charcoal at the base of the pit. And then the rest of the pit stays open for a while and is gradually filled in with junk, debris, which included a tablet of red ochre, a bifacial core and all sorts of interesting stuff. It was just a depression in the ground that acted as an artefact trap. When we dated the charcoal that formed the rim of the base of the pit, it was exactly the same age as the charcoal from the level at which the pit was dug. Here we have a little pit. We never quite worked out why it was dug, but it was dug quickly; it was partially infilled; and it then accumulated other objects.
There have been some quite remarkable examples of features found and I’m thinking at the Roonka Flat site on the Lower Murray. Excavation of the upper levels of this dune site started to reveal some unusual features that were very carefully exposed in the sandy sediments. Gradually, it became clear that this was a line of tractor treads going diagonally across the trench. In this case, the local farmer had graded the top of the dune for a lucerne paddock and had triggered a whole phase of erosion. A few years later, the South Australian Museum had come along to excavate the site and salvage the burials. In the course of that they excavated an area where drift sand had moved over the original surface that had been bulldozed, and they found these wonderful tractor treads weaving across the site.
Other examples include the Willandra Lakes footprints, where footprints have been pressed into soft calcareous earth and then hardened and then buried by dune sand. Certainly in Namibia, in muds, in clays, there are quite extensive animal trackways in the Namid Desert where they’ve got elephants moving across the landscape.
At Lake Callabonna in South Australia there are also diprotodon trackways. The diprotodon tracks appear as just a depression in section, but in plan, of course, you can see you’ve got entire footprints. So footprints and spoor are a particular feature to keep an eye out for.
I will stress here that features are very important. If you’re looking at post-holes, they might give you a guide to whether people were building structures on the site. Features are a disturbance, but features are also an opportunity to look at natural events.
This is Danebury, an Iron Age hill fort in Britain excavated by Barry Cunliffe. It’s a really good example of the care and attention given to features, the internal detail of these features, how they’ve silted up and whether they have secondary features within them. These are storage pits dig within the actual hill fort itself. We can see a lot of the internal details in these features - the actual ditch but also the material that’s been dug out of the ditch as an embankment and all the other features dug into a feature.
ROBERT PATON: Just taking up the point of excavating features across the site to expose them across the site, how important it is at our work that we did recently in Tasmania. The top 20 centimetres of the site had been ploughed and we were hoping to use OSL [optically stimulated luminescence] dating to work out the age of the site. So we had to work out what had been ploughed and what hadn’t. We were fortunate enough to be able to just move slowly and expose the entire ploughed area, every furrow showed up so beautifully.
But in section it would have been almost useless because we would have probably gone down another 20 centimetres before realising that we’d gone into those sediments which were undisturbed below, which was, of course, another 10,000 years. So your point about not just doing features in section I think is highly important.
MIKE SMITH: I have two other examples of features that I think are worth drawing attention to. One is at the basal levels of the Malakunanja site. We had quite a rich horizon of stone artefacts at 50,000 to 55,000 years. The real questions were about whether the artefacts were in situ or could have moved down from higher levels. I don’t think they did, but we were lucky enough that this particular trench clipped a small feature pit, which is quite an unusual feature, that was obviously dug in the sandy floor of the rock shelter. And then it’s almost as if all the artefacts, all the sharp pieces of stone, on the surface of the site had been scooped into the pit. There were large pieces of ochre, one-kilogram pieces of hematite, large numbers of stone artifact and a jumble of fragments of grindstones that matched pieces elsewhere in the site. It was just one of those little events, but the importance of it was that it was intact. An intact feature isn’t something that has been moved down a sequence or disturbed in some way and it was constructed some time before 45,000 years. That at least gave us a point in stratigraphic space and time where we were confident that we had some integrity to the deposit.
My second example concerns Cuddie Springs. Cuddie Springs is quite a remarkable Pleistocene site in an old swamp system in New South Wales. Judy Field dug a very large open area excavation. The main Pleistocene occupation unit is very rich in stone artefacts, so we’re looking at tens of square metres of exposure. But there were no features. There was no apparent concentrations of artefacts; there were no hearths; there were no napping areas. I think this is quite informative about the processes that operated on that site in terms of the movement of material within these sort of fluvial and lacustrine swamp systems. We’ve had perhaps sufficient vertical movement, maybe not over large distances, but sufficient movement of swelling and cracking of clays to actually destroy whatever features are there. Maybe you even had removal of fine material as well, but basically it looks more like a lag of artefacts. I’m not talking about a stone bed, I’m talking about the material underneath it. It still looks more like a lag of artefacts than it looks like an occupation surface. If you clear 100 square metres of surface in a heavily occupied site, you should find some features - some ovens, concentrations of burnt clay, napping areas - and you don’t. So that in itself is quite informative.
This is a stratigraphic diagram from the rear of Puritjarra rock shelter. It shows a number of things. First of all, as we move along layer one, you can see that, as you get toward the rear of the rock shelter here and here, you get a lot more local accumulation of rocks fall as the wall of the rock shelter weathers. It’s very much a lateral variation rather than anything else. You can see the small hearths and fireplaces that were used as sleeping fires towards the rear of the rock shelter.
The feature I really want to show you here is this pit here, and also shown in plan here. It was a pit that was dug under the rim of a large boulder, ultimately this boulder here, and it was rapidly backfilled as it was dug. So we have charcoal samples from here and a clod of earth that’s fallen in, both of which date to this level up here, the level from which the pit is dug. It obviously stayed open for some time. We have a little fill line here of rocks and above this accumulated a whole range of artefacts, including a core, a piece of ochre and other artefacts, and other big pieces of charcoal.