When the CFO arrived at Settlement A, there were four funerals under way in the immediate region: one at A itself, one at homeland A10 and two at other large predominantly non-Yolngu settlements in the wider region (G and H). The CFO had difficulty recruiting local CIs for A and C. Using lists of people supplied to him by the council offices of A and C, he found and trained one CC for Community A and two for Community C, and two potential CCs for the homelands attached to Community A. He and the CCs tried and failed to recruit CIs for the training. It had been estimated back at the Census Management Unit (CMU) that eight CCs and 21 CIs would be needed for the count in these two communities and their associated homelands, so this was not an auspicious start.

Having got the count under way in sections of the main settlements that were not affected directly by the funeral, the CFO set off down the track—accompanied by myself and the Assistant CFO—to survey the situation in some of the homelands. We drove into A7 (a three-dwelling homeland) towards evening. It was empty—everyone was at the funeral at A. The next day we called in at A8, which has a usual resident population of about 80. It too was deserted—everyone was at the large funeral happening at nearby A10.^[3]

It was decided to return to these communities after the funerals were over and people had returned. The CFO had been warned before leaving A—by staff of local organisations and by the CCs he had recruited—that these homelands would be empty, or nearly so. Because people know how everyone is related to everyone else, they can predict with a fair degree of certainty which people will be attending which funerals. At this stage, the CFO—who had not previously worked in the Yolngu area, apart from undertaking the Community Housing and Infrastructure Needs Survey (CHINS)—had not fully appreciated the potential impact of funerals on population movement, and wanted to see for himself.

We then called in at A9. Not many people were there—some were at the funerals at A and A10. The daughter of the community leader was seriously ill in hospital and her parents had just returned from seeing her there. Nevertheless, the community leader and his wife were happy to be recruited as the CC and CI for A9 and to conduct the count once everyone had returned from the funerals. They also volunteered to enumerate A10 once the funeral there was finished. The CFO arranged to come back in a few days to conduct the training, and we set off further down the track in an optimistic frame of mind.

We drove on to the largest homeland in the area, C3, where the usual resident population is in the region of 160. The plan was to recruit another CC and some CIs here to cover C3, C4 and C5 (another large homeland), since these three are close to each other and their inhabitants are closely interrelated. The CFO was successful in recruiting two CCs (one of whom had been a CI during the 2001 Census) and three CIs. After their training, however, as the CCs and CIs began going from house to house, it became evident that at least one-third of the population of C3 was away, at four different funerals (the ones at A, G, H and A10), or visiting relatives, predominantly at A, and several people were about to leave for the final part of the funeral at A10. In two cases, the entire household was away and their dwellings stood empty. It was decided to do a first count here, go on to C4 and C5, then return to C3, hopefully to catch people as they returned from the various funerals. In the meantime, all those who were away were put down as ‘persons temporarily absent’ (PTA), including on the forms for the empty dwellings. There were at least 20 ‘visitors’ present—people visiting their kin from other communities (predominantly C2, C4, C5 and the settlements of A, C, B and G).^[4] These were counted as visitors in the main section of the form. The C3 team was about to go to C4 and C5, and there was a possibility that some of the visitors would be returning to these communities at the same time, but the CFO and the C3 team were confident that no double-counting would result because everyone would be able to keep track of who had already been counted as visitors at C3.

I stayed with the C3 team, and drove them on to C4 and C5 when they had finished their first pass at C3. The CFO and his Assistant returned to A to try to get the count going in the other A and C homelands, and to offer support to the single remaining CC at A, who at this point was endeavouring to enumerate this sizeable community on her own. Back at A, the CFO discovered that the entire populations of A3 and A4 were at A for the funeral there. They were enumerated at A ‘as if’ they were in their dwellings in A3 and A4. I will return to the reason why this happened below.