Chapter 6—
The Rent Histories of Estates and Houses

Delian Apollo owned about twenty properties, called

in the accounts and conventionally referred to in scholarship as "estates," which were located on Delos itself, on Rheneia, and, after 237 B.C. , on Mykonos. Most of them were farms, producing grain, livestock, figs, and wine, but unfortunately (except for the Mykonian holdings) no olive oil. For most of the years of independence, the estates were leased after secret bidding for ten-year periods. Except for a few gaps, the most serious of which covers 240–220 B.C. , records of the rents of the estates are fully preserved. Every tenth year, before the next rental period, the hieropoioi recorded detailed inventories of the estates' capital equipment, including plantings of vines and fruit trees, which permit some inferences about the strategies of exploitation followed on each estate.

This unusually rich material offers a counterpoint to the price data studied in the previous chapter. Land rents can be compared to the movement of prices for agricultural goods, an enterprise impossible virtually everywhere else in the Hellenistic world except Egypt. While Delos was not part of a large-scale, pan-Aegean price-setting market, there was clearly a local market for agricultural products, and the impact of that market can be read in changes in the rents for the estates; we can sometimes even make inferences, tentative though they may be, about conditions in the neighboring Kyklades. We can explore the differing impact of specialization in viticulture, stock raising, and cereal culture in rents over time, as local economic conditions changed. Even some of the noneconomic factors that entered into decisions by the Delian elite to rent Apollo's estates can be elucidated.

The accounts also preserve records of a series of "sacred houses" (

) owned by Apollo and rented on five-year leases by private persons. Unfortunately, the rents of these buildings, although fairly well-preserved

---

across the years of Delian independence, cannot be subjected as easily and directly as the estate rents to economic analysis. Nevertheless, they provide important evidence about the economic situation in the city of Delos itself as opposed to the countryside, and open a window on changes in the urban economic situation, which did not always move in conjunction with the rural one.

The Origin and Character of the Estates

We can only surmise how the first estates came into Apollo's possession. In 523 B.C. , Polykrates of Samos captured Rheneia and gave it to Apollo, and John Kent has reasonably conjectured that this gift was the origin of some or all of the god's Rheneian estates. Those on Delos may likewise have been gifts or dedications; Kent's suggestion that Hippodromos was a gift from the Athenian general Nikias in the fifth century seems to have won general acceptance.^[1] Several more estates seem to have come into Apollo's possession as a result of seizure of property of Delians convicted of sacrilege in 375 B.C. ; the origin of yet other estates, which first appear in the accounts between 297 and 282 B.C. , is treated below.^[2] However Apollo came to own them, the estates constituted an important part of his wealth, as shown by the scrupulous regulations that govern their rental (ID 503, the

, on which further below).

For the amounts of the rents I generally follow the figures given by Kent,^[3] except that for several estates in the early period I have omitted rents reconstructed on the assumption that some missing rentals were renewed without competitive bidding by the application of a 10 percent increase, the

. Kent is usually probably right in his reconstructions, but since they do not affect the rent history I have preferred to set them aside. I have also used the original of two rents when a renter defaulted in the middle of a lease and the estate was rerented. Sometimes these estates fetched considerably less in rerental, but not, I think, for economic reasons: the new renters were no doubt taking advantage of an unexpectedly vacant estate to pick it up at a bargain rate.

Beginning with the year 299 B.C. , table 6.1 shows rents for each ten-

[1] Thuc. 1.13.6, 3.104.2 for Polykrates, with Kent, 245; Plut. Nik. 3.7, Thuc. 3.104.6, Kent, 255–56, Brunet, 31–32, Isager-Skydsgaard, 183, for Hippodromos.

[2] See ID 98B24–52 for the confiscations; Kent, 256–58, 286–89; Tréheux, "Dernières Années," 1016 n.2; Brunet, 31–35, 49–64.

[3] Kent, 303–4, still the standard treatment. There is, however, much to correct; see Jacques Tréheux, Rev. Arch. 31/32 (1948): 1008–32, esp. 1011–22; Brunet, passim; Heichelheim, Wirt. Sch., 82–83; Larsen, 402–7, esp. 403, 406. And see Appendix IV, pp. 309–38, below.

---

---

year rental period.^[4] The figure given under 299 B.C. was that submitted by the winning bidder (estates were leased by sealed bid; see further below) in 300 B.C. for a lease beginning in Lenaion 299 B.C. and running through Posideon 290 B.C. This term was the result of the implementation of a new rental contract, the

, which called for ten-year rental periods.^[5]

Before 300 B.C. , rental periods were very irregular. In the fall of 314 B.C. , when Antigonos Monophthalmos and his son Demetrios freed the island from Athens, the Delians cancelled leases then in force and replaced them with new four-year leases that ran until 310 B.C. Four Delians were apparently allowed to renew their leases by paying 10 percent more rent (the

); the other estates, which had probably been in Athenian hands, now fell to Delians.^[6] Over the next few years, leases were renewed several times. In 310, the estates were put up for five years (IG XI 2.143B1–2), but the lease period that began in 305 B.C. cannot have lasted longer than three years, for the payments in IG XI 2.144A9–17 (304 B.C. ) are different from those of IG XI 2.146A9–12 of 301 B.C. There were therefore four separate leasing periods between 315 and 300 B.C. Table 6.1 dates the earlier ones by the dates of the inscriptions to which we owe our knowledge of the rents, and not by the dates of the leases:

1. Four-year lease, 314–310 B.C.

2. Five-year lease, 310–305 B.C.

3. One-, two-, or three- year lease, 305–304/3/2 B.C.

4. Two-to four-year lease, 304/3/2–300 B.C.

For purposes of analysis, it is useful to try to sort the estates into groups that share important characteristics. Groupings based on the primary products of each estate would clearly be interesting for economic analysis. In the years when the estates were put up for lease, the hieropoioi made a

[4] For discussion of individual figures, see Appendix IV, pp. 309–38, below. Unfortunately, in some respects the data collected there are provisional. Several important new discoveries await publication: the join of ID 456 + 440 (photograph at J. Tréheux, BCH 109 [1985]: 486, fig. 1); a new reading of 356bis; and the join of 374B with a new frag. G 761, for all of which see J. Tréheux, BCH 110 (1986): 430–31 (Brunet, 45). For ID 452 + 467, see M. Brunet, BCH 114 (1990): 678–79, and for the date G. Reger, Hesperia 63 (1994): 105–10.

[5] On the date, see Appendix I.

---

thorough inventory of the capital goods of each estate. The results were incorporated into the accounts as part of the contract between Apollo and the renters. Using these inventories (especially the very well preserved one for 250 B.C. in IG XI 2.287A134–74), Kent posits four estate groups based on product. His first group consists of eight estates that produced grapes, grain, and livestock. The second group comprises three estates primarily devoted to livestock, with some vines. The four estates of his third group lacked vines altogether, and the products of the five estates of the fourth group could not be identified. Kent's groupings were especially interesting in view of his hypothesis that declining wine prices in the second half of the third century depressed the rents for estates that were dependent on wine.^[7]

Unfortunately, close inspection of the evidence for estate products fails to support Kent's groupings. His third group indeed includes only estates that clearly lacked vines—Hippodromos, Kerameion, Leimon, and Soloe-Korakia—but one of them, Kerameion, is by Kent's own description "a potter's establishment."^[8] It clearly should be eliminated from the group, whose other members are agricultural enterprises. Two of the three estates in Kent's second group, devoted mainly to livestock but with some vines, each have a

and one, Phoinikes, has an , both of which suggest grain production—said by Kent to be characteristic of his first group.^[9] Moreover, Rhamnoi, although placed in this group, supported 1,978 vines, almost three times as many as the 700 that qualify Nikou Khoros for placement in Kent's first group. One of the estates Kent puts in his fourth group, whose products are said to be unidentified, is Skitoneia, which in fact had 629 vines but no or , which would seem to make it preeminently dependent on wine production.

In fact, capital equipment in the form of buildings may in general offer a poor guide to the products of the estates. Every estate but Dionysios had a

, implying that they all produced at least some grain; and Dionysios in fact had an , which means that it produced "chaff" in some manner. Two estates known to have produced barley in the late fourth century—Hippodromos and Soloe-Korakia—both lacked -

[7] Kent, 310, 299–301; 309–13.

[8] Kent, 254, with n. 25.

[9] The products of the third estate in the group, Epistheneia, known to Kent only from the very fragmentary ID 373A8–13 (cf. Kent, BCH 68 [1939]: 242) and 467.1–4 (cf. Kent, BCH 68 [1939]: 245), have now become clear thanks to Tréheux's join of ID 452 and 467 (BCH 110 [1986]: 430–31), published by M. Brunet, BCH 114 (1990): 678–79.

---

. The absence of on Limnai and Skitoneia, then, cannot prove that neither estate supported herding.^[10]

There is, however, one factor that can be used unambiguously to sort the estates, and that is the presence or absence of vines. IG XI 2.287 proves conclusively that—at least in 250 B.C. —Hippodromos, Leimon, and Soloe-Korakia had no vines at all. The three estates therefore make a clear group, which I shall call Group I. Of the remaining sixteen estates, twelve are known to have had vines, but in considerably different numbers. Three estates supported more than 1,900 vines: Pyrgoi with 2,250, Kharoneia with 2,187, and Rhamnoi with 1,978. These I call Group IIA. Four estates, Group IIB, had between 1,550 and 1,000 vines: Porthmos with 1,535, Limnai with 1,514, Panormos with 1,298, and Dionysios with 1,056. A final set, Group IIC, consists of estates with 700 or fewer vines: Nikou Khoros with 700, Skitoneia with 629, Phoinikes with 596, Khareteia with 560, and Epistheneia with 487.^[11] This sorting will allow a close test of Kent's hypothesis that declining wine prices depressed estate rents after the mid third century. (The estates are arranged by group in table 6.1.) For our purposes I have calculated indexed values for each estate's rents and then derived an aggregated average rent for each group in each rental period. The aggregated data appear in table 6.2.^[12] It is these aggregated data that we shall use in our analysis. The same data are displayed graphically in figure 6.1.

Figure 6.1 offers graphic illustration of the decline rents underwent in 290 B.C. ; as we shall see, analysis confirms the reality of this change. Another, more modest adjustment can be detected around 220 B.C. Various hypotheses have been suggested to explain these changes, including Kent's view, already mentioned, that declining wine prices after 250 B.C. were to blame for the loss in value of the estates after 220 B.C. These earlier views are tested below. First, however, I would like to explore another approach, which, strangely enough, has not been tried. As we saw in chapters 4 and 5, we have series of prices for several important agricultural commodities,

[10] On the terms, see Brunet, 180–81, 187–98; Robin Osborne, BSA 80 (1985): 121; Vallois, 1.213; Kent, 292–301; and see generally, Jan Pecirka^[*] in Problèmes de la terre en Grèce ancienne (Paris, 1973), 137–40.

[11] For three remaining estates—Akra Delos, Lykoneion, and Sosimakheia—no inventories survive. On Phytalia, see below. The evidence is conveniently set out at Philippe Bruneau and Philippe Fraisse, BCH 105 (1981): 141–44, with a table at 143 (missing however Epistheneia).

[12] For full data, see Appendix IV.

---

including barley, olive oil, firewood, and pigs. Since some of these goods (especially barley) were produced on many, if not most, of the estates, nothing is more natural than to see whether the rents renters paid were determined by the prices of agricultural commodities.

Commodity Prices and Estate Rents, ca. 280–169 B.C.

The estates produced a variety of agricultural commodities—grain, wine, livestock, figs—of whose local prices the renters must have been perfectly aware. Since much of the Delian demand for these goods was satisfied by local Delian or neighboring Kykladic production, renters who wanted to dispose of products on the local market would have had to take these prices

---

Figure 6.1.
Average Aggregated Indexed Rents of Apollo's Estates, by Group, 314-179 B.C.
Note: No data available for years 248-220, or for Group II for 198-180.

into account when reckoning the level at which they were willing to bid for an estate. Once the hiera syngraphe had eliminated the less wealthy and more speculative bidders, whose interest in the estates may have had a considerable social component (as we shall see below), rents ought to have centered increasingly on economic considerations.^[13]

Unfortunately, several difficulties stand in the way of testing this hypothesis. For the most important agricultural commodity, grain, we possess only a handful of prices (see table 6.3), and only a few of these date to years when the estates were up for rental. The situation with olive oil prices is considerably better, since we have a very long series of prices for that commodity, and many that date to years when estates were bid out; but unfortunately not a single estate except on Mykonos had any olives. Firewood

[13] On the "commercial" character of these enterprises, see Claude Vial in L'Origine des richesses dépensées dans la ville antique (Aix-en-Provence, 1985), 47–53; Vial, 317–38; and Robin Osborne, Chiron 18 (1988): 302–304, who writes that "the pattern of agriculture on the temple estates made an openended commercial attitude to farming possible: the turn-over of estate leases makes it look as if some, at least, [of the renters] took advantage of that possibility" (302) and observes that on Delos "agricultural activity seems to have become devoid of social value" (304). For another approach to the relation between grain prices and rents, see Robin Osborne, Classical Landscape with Figures (London, 1987), 46–47.

---

prices are likewise abundant, but it may fairly be doubted whether the collection and sale of firewood was of any real economic importance to estates without copses. Sheep prices are far too rare to be of any help, whereas the long series of pig prices presents its own problems, as we shall see.

The earliest years of Delian independence present even worse problems. While we can document a relation between commodity prices and rents for each rental period from 280 to 250 and again from 220 to 180 B.C. , there are only one usable oil price and one usable pig price for the period of high rents (314–291 B.C. ). Scholars have long argued for a connection between these rents and prices, but this view cannot be tested for want of data. I shall therefore postpone treatment of this period and concentrate here on the years after 280 B.C.

For this analysis, I have compared rents with prices for goods from the year in which new bids were submitted, from the closest year immediately preceding, or, in a few cases, when no other data were available and prices were relatively steady over time, from the immediately following year. This procedure assumes that bidders were most likely to be affected by prices current when they offered bids. There may, of course, have been fluctuations unknown to us that affected rents, but we have no way to control for these. In any case, there are some interesting relationships between rents and prices that help to alleviate worries about the potential inadequacy of the data.

Grain Prices and Rents

As grain, and especially barley, was the staple food in the Kyklades, it is reasonable to begin with a comparison of grain prices with estate rents. Unfortunately, grain prices are scarce for Delos, and those for wheat are too rare to use here; the few barley prices are given in table 6.3 in the form of cost of processed alphita. Any comparison between alphita prices and rents must be treated with extreme caution, because so few comparable data points are available (only six for Group I and five for Group II estates), but the results are extremely interesting. For Group I estates, the correlation is almost perfect. To the extent it is possible to test, estate rents move in exact conjunction with alphita prices (fig. 6.2; table 6.4).

For Group II estates, the connection is considerably less definitive (fig. 6.3, table 6.5). Although Groups IIA and IIB show a correlation—47 and 49 percent respectively of the variation in rents can be explained by changes in barley prices—the relation is not very significant (in both cases

---

Figure 6.2.
Movement of Group I Estate Rents in Relation to alphita Prices, 279-179 B.C.

---

Figure 6.3.
Movement of Group II Estate Rents in Relation to alphita Prices, 279-179 B.C.

---

---

the two-tailed significance is greater than 0.1). Group IIC is another matter; 61 percent of its rent variation can be explained by variation in barley prices, and the correlation is significant.

Based though they are only a handful of data, these results are striking and important. There is a clear relationship between the absence of vines and the degree to which estate rents are linked to alphita prices; this link becomes almost absolute for Group I estates, which lacked vines entirely. No one can doubt that the renters of these estates had a clear idea of the local market value of the goods the estates produced, and that they calculated their bids accordingly. Furthermore, the increasingly strong correlation between rent and barley prices as one moves from Group IIA estates through Groups IIB and IIC to Group I casts doubt on the supposition, offered by Kent, Brunet, and others,^[14] that herding was the main income-producing activity of estates without vines. In fact, it would seem that barley culture, not herding, was their chief business (although there was doubtless some herding on all of them, as they all had sheep shelters).^[15] As I show below, there is no correlation of rents with pig prices except perhaps for Group I estates after ca. 220 B.C. , which may represent a change in patterns of exploitation. But until then, and certainly always for Group IIC estates, barley was likely to have been their most important agricultural product.

Olive Oil Prices and Rents

Even though none of the estates had any olives, some of their rents do show striking correlations with oil prices. The results for Group I estates are set out in figure 6.4 and table 6.6. Figure 6.4 makes it quite clear that, with the exception of the decade between 280 and 270 B.C. , oil prices and rents for Group I estates were tightly linked. Indeed, 74 percent of the variation in rents can be explained simply by variation in oil prices (table 6.6). While we must be cautious about the implications of the single datum point available for before 290, there can be no doubt that from about 270 B.C. on, changes in the price of oil predict the changes in rent levels for these estates very well.

The story becomes more complicated in Group II estates. Figure 6.5 appears to reflect relations between oil prices and rents very like those iden-

[14] Kent, 301 (no. XI, Hippodromos; XII, Leimon; XIV, Soloe-Korakia), 309–10; Brunet, 140–42; Bruno Cavagnola, Istituto lombardo, rendiconti, Class. di lett. e sc. mor. e. stor. 107 (1973): 513; Isager-Skydsgaard, 196–98 (less precise about which estates).

[15] Kent, 300.

---

Figure 6.4.
Correlation of Olive Oil Prices with Group I Estate Rents, 304-179 B.C.

tified for Group I estates; very interesting again, but difficult to interpret because of the extreme paucity of data points, are the apparent correlations between high prices and high rents (304–303 B.C. ) and the rise in rents in the decade 280–270 while oil prices fell. For the decades after 270 B.C. ,

---

Figure 6.5.
Correlation of Olive Oil Prices with Group II Estate Rents, 304-179 B.C.

however, the subgroups of Group II, although showing at least some correlation with oil prices, diverge from one another and from Group I in a very interesting way: the degree of correlation between oil prices and rents varies inversely according to the number of vines (table 6.7). Group IIA estates, which depended heavily on wine production, show poor correlation with oil prices. Only 23 percent of rent variation can be accounted for by changes in oil prices, the significance of the correlation is poor, and the amount of auto-correlation as measured by the Durbin-Watson is unacceptable. An attempt to correct for the latter eliminates all significance from the relationship.^[16] In other words, decisions about the level of rents renters were willing to pay for Group IIA estates were essentially independent of oil-price levels. Rents for Group IIB and IIC estates, on the other hand (table 6.7 B, C), moved in conjunction with oil prices. Half of the variation in Group IIB estates, with a moderate number of vines, is linked to oil-price change; and fully 66 percent of the variation in Group IIC can be attributed to oil-price changes. This last figure approaches the result for Group I estates.

The first two patterns apparent for Group I estates seem reasonable.

[16] As seen when the Cochrane-Orcutt procedure is applied, giving a two-tailed significance of 0.501 (!) and an adjusted R-squared of only 0.23.

---

---

High oil and high estate prices might be linked before 290 B.C. , as many scholars have thought (but see below), but the data points are so few that it is very difficult to say. That rents and oil prices should move in different directions in the 270s occasions no surprise, since the estates had no olives and therefore their rents should not have been affected by oil prices.

But the results for the years after roughly 270 B.C. are striking. The pattern virtually reproduces the results for barley: the fewer the vines, the greater the link between oil prices and rents. These results seem strikingly to confirm D. Rathbone's view that "agricultural prices in general, including that of olive oil, are likely to have followed similar price trends in the long term."^[17] But the linkage between oil and grain prices implied by the correlation of both with rents is itself puzzling. No clear mechanism accounts for it. The natural history of the two crops is very different (grain was planted in the fall/winter and harvested in the spring/summer; olives were harvested in the fall); their demands on labor fall at different times of year and at different levels; their rainfall requirements are strikingly different as regards both amount and timing. Most important, they cannot have been substitute goods, like wheat and barley. When wheat was scarce, consumers turned to barley, and the rising demand drove the price up. But it is hard to see why consumers unable to get barley would turn to oil, especially as oil was a relatively expensive product in antiquity anyway.

This correlation presents many questions, and it is impossible to offer a definitive solution. Reflection on the local situation on Delos and the neighboring Kyklades, however, may suggest a tentative hypothesis. In chapter 5, I argued that, after a period of dependence on extra-local sources for olive oil in the late fourth and early third centuries, Delians and their immediate neighbors planted olives and began to supply local needs from local sources. (The failure of the hieropoioi to plant olives on Apollo's estates must be attributed to reluctance to make the long-term investment on rented property.) This development reduced and stabilized oil prices. It would also have had an effect on the local price of land, for once the olives began to produce, their yields would have become part of the income-producing capital of the property they grew on. Oil prices would then generally have become a factor in the equation by which land values, and so rents, were set. This linkage would be still stronger if it was customary in the Kyklades—as would seem very likely—not to plant olives in dense orchards but to scatter them about the landscape with enough space in between to grow grains. Such a mechanism could explain why even rents for

[17] D. Rathbone in Eighth International Economic History Congress, Budapest 1982, ser. B12 (Budapest, 1982), 48.

---

Apollo's estates, which had no olives, followed oil prices. The power of this explanation must not be exaggerated. Unlike the correlation between rents and barley prices, that between rents and olive oil prices breaks down on occasion. This indicates that the tie is not direct, as I have argued for barley, but indirect, perhaps through the mechanism suggested here. Indeed, it seems likely that, rather than trying to explain rents as a dependent function of oil prices, we should seek a larger explanation to encompass both. My suggestion that planting olives helped to stabilize land prices is such a larger explanation, but, as we shall see in chapter 7, there are still other, more general, factors that need to be considered.

Firewood Prices and Rents

Not surprisingly, goods that moved in a different economic sector, with different constraints on production, show no price correlation with the estates: firewood prices fail entirely as predictors of rents. Firewood, of course, was produced in a very different way from grain or oil; and the estates entirely lacked copses where wood would have been gathered. It is therefore no surprise that firewood prices do not correlate with rents in any significant way.

Livestock Prices and Rents

Livestock would seem to have been a different matter. The estate inventories assure us that many estates were equipped to run livestock, and Kent and others have supposed that herding was an important generator of income for many estates, including especially those of Group I. It is, however, difficult to test for a connection between livestock prices and rents. There can be no doubt that the majority of the estates'

were sheep and goats, but prices for these animals are too few to try testing for a connection. Pig prices, of course, are abundant, but I hesitate to use them as a substitute. As I argued in chapter 5, pig prices were tied to firewood prices for reasons that seem to stem from the methods of raising hogs. This means that pig prices were set, at least in part, by mechanisms that had little or no impact on rents. Moreover, pigs and sheep or goats cannot be regarded as substitute goods. Whereas hogs provided only meat (and, of course, sacrificial victims), sheep and goats were raised primarily for milk and (in the case of sheep) wool; only secondarily did they themselves serve as food. Finally, pig prices were extremely volatile; they might fluctuate by 40 percent, 50 percent, or more from month to month. Such frequent and extreme changes may have made their prices useless for gauging the profitability of an estate.

---

Figure 6.6.
Correlation of Pig Prices with Group I Estate Rents, 301-179 B.C.

Yet pigs were surely raised on at least some estates (see chapter 5), and although over the whole period of independence, pig prices do no better than firewood as predictors of rents, they may have left a trace in the rents of Group I estates. Unlike the estates of Group II, these show a slight but definite rise in rent in the later third and second centuries. This period after ca. 220 B.C. is exactly the time when mean annual pig prices (following firewood) climbed to a new and permanently higher level (fig. 6.6). Although it is impossible to be sure, it may be that the renters of these estates, which were devoid of vines and had previously been devoted largely to cereal culture, began, in response to the rising price of pigs, to run more hogs, and thus to reckon the value of these enterprises in part in terms of pig prices. If this is right, it points to yet another shift in the economy of Delos in the late third century, to which I shall return in chapter 7.^[18]

I cannot leave this topic without discussing one other issue, and that is the character of Phytalia, an estate that was regarded as an orchard until the recent join of ID 452 and 467 provided the first inventory for it, showing that it was devoid of any capital equipment at all.^[19] Its rents show no

[18] See the somewhat similar view expressed at Isager-Skydsgaard, 196–98.

[19] W. Déonna, La Vie privée des Déliens (Paris, 1948), 96, and BCH 70 (1946): 160; Kent, 254 n. 25; Vial, 323 (a jardin). For the join, see n. 9; for the inventory, see II. 24–25.

---

Figure 6.7.
Olive Oil Prices and Rents for Phytalia, 290-179 B.C.

important correlation with barley prices and a correlation of about 40 percent with oil prices (fig. 6.7, table 6.8). The absence of any correlation with barley prices makes it very unlikely that Phytalia was, as Brunet suggests,

devoted to barley culture,^[20] and the moderate correlation with oil (roughly between the responses of Group IIB and IIC) reflects only the general impact of oil prices on the rent of land on Delos. What was Phytalia then? It may be that here, and here only, we have a true ranch: a tract of land without capital equipment, used by renters to run sheep, goats, or other livestock, which were housed elsewhere.

The results of this section are very important, for despite the few data points on which they rest, they are remarkably consistent: the rents fetched by agricultural land on Delos were linked to and in part determined by the price levels of important agricultural commodities. This linkage expressed itself most strongly in the estates of Groups I and IIC, and to a lesser extent Group IIB, which had no, or relatively few, vines. Barley culture provided the direct connection, since these estates in fact produced barley. The tie with olive oil, at first sight surprising, can probably be explained by the

[20] Brunet, 146; cf. BCH 114 (1990): 679.

---

role oil played on land in private possession on Delos and its nearest neighbors.

These two basic staples, barley and olive oil, which were locally produced and locally consumed, were thus closely linked in the local economy, and rents were tied to them. This sheds welcome light on a very important aspect of the local economy. In contrast, growing grapes and running livestock seem to have taken place in a different sphere. Livestock have left surprisingly little impact in our data, although the picture might be quite different if we had a set of prices for sheep or goats comparable to those for pigs. Viticulture, of course, was aimed mostly at wine production, and we shall have to test Kent's idea that declining wine prices after 250 B.C. depressed the values of Group II estates. Whether we can say a "universal price-setting market" for wine existed will depend in part on the results of that investigation.

Rent History until 290 B.C. and Kent's Wine Hypothesis

Two important phenomena in the rent histories of the estates have long been noticed: the drastic collapse of rents in 290 B.C. , and a slower decline after 220 B.C. , which Kent attributes to depressed wine prices.^[21] Both of these phenomena can be confirmed by statistical analysis, which

[21] Kent, 302, Larsen, 401–7, Heichelheim, Wirt. Schw., 82–83, Gustave Glotz, Journal des Savants 11 (1913): 19–20. Kent, 310, 299–301; 309–13.

---

however also reveals new and important differences among the groups of estates.

The rent history of Group I estates responds very nicely to a model that attributes almost all of the change in their rents to a turnaround in 290 B.C. This model explains fully 72 percent of the variation in rents and indicates an extremely strong response to the watershed; rents decline by almost 106 units (table 6.9). There is, however, no improvement in explanation if the model is modified by adding a dummy variable for the period after 220 B.C. This is an important result, since it seems to confirm Kent's view: these estates without any vines at all are not affected by his proposed decline in wine prices over the second half of the third century.

For Group IIA estates the best model incorporates dummy variables for both 290 B.C. and for Kent's wine factor, taken as a watershed at 220 B.C. This model accounts for fully 88 percent of the variation in rents with good statistics (the Durbin-Watson is marginal but acceptable). It is interesting that the coefficient for 290 B.C. represents a much more modest response than that for Group I estates (table 6.10).

Group IIB estates, which had 1,550-1,000 vines, follow a pattern similar to that of Group IIA estates (table 6.11). Eighty-three percent of their rent variation can be attributed to the two factors, 290 B.C. and Kent's wine hypothesis. The response of rents to the watershed years, however, is rather different from those for the IIA estates. The decline in 290 B.C. amounts to 61 units, or 2.25 times greater than for IIA estates, while the decline in 220 B.C. of 48 units is essentially the same as that for IIA estates.

The same pattern obtains for Group IIC estates (table 6.12). Again, the

---

model accounts for 87 percent of the rent variation, and the coefficients that represent the response to each factor are close to those of the Group IIB estates. On the evidence of these results, Groups IIB and IIC clearly belong together. Group IIA estates stand apart in their more moderate response to the change in 290 B.C.

---

Rent Levels, 314-290 B.C.

The results show unequivocally that important adjustments in estate rent levels occurred on Delos for all estates around 290 B.C. Before 290, rents were extraordinarily high, and they collapsed in that year; they never again reached levels even approaching the dizzying heights of the late fourth century. Several hypotheses have been offered to account for these extraordinarily high rents. Fritz Heichelheim and others have pointed to the high prices for agricultural commodities for the same period and suggested that these prices, attributed to the great demand for "Greek goods" by the new settlers in the East, drove up rents.^[22] The close tie found above between agricultural prices and estate rents would seem to support this view, and indeed it must be admitted that oil prices at least were very high in this period.

There are, however, some problems with this view. Consider first the link between prices and rents. Group I estates show the tightest correlation with alphita prices (fig. 6.2). This correlation can be used to work out a very rough formula for calculating putative barley prices from rents (multiplying the rent by about 0.0422 gives an alphita price in dr). By this formula, barley prices in 314-290 B.C. ought to have ranged from a low of 6.9 dr/med in 310 to a high of 12.4 dr/med in 301 B.C. ; the average price would have been 8.6 and the median 7.3 dr/med (cf. table 6.3). These prices

[22] Heichelheim, Wirt. Sch., 55–56; Larsen, 380; Rostovtzeff, 165–66; cf. also H. Michell, Canadian Journal of Economics and Political Science 12 (1946): 3–7.

---

are extraordinarily high; indeed, the highest alphita price actually attested on Delos is only 5 dr/med, in a year generally considered to have seen a severe shortage.^[23] Such prices; persisting over twenty-five years, would have represented a disaster of enormous proportions, far more serious than the shortages that beset Athens in the 330s and 320s.^[24] It is hard to believe that such a situation would have left no traces in the sources at all.

Furthermore, it is difficult to understand how exactly the high commodity prices are supposed to have raised land rents. Barley was produced in abundance in the Greek East, sown certainly from the first day the colonists arrived (see App. Syr. 1.1). It is hard to imagine an eastern demand for barley lasting twenty-five years that could drive up prices at Delos to as high as three or four times the typical price level in the mid third century. Oil is another matter, since trees require many years to produce new crops; but here is a puzzle too. The Delian estates produced no oil. The tie elucidated above between oil prices and rents works well in a market of relatively stable oil prices, in which olive-bearing land is priced in the same local market as land without olives, but these relations would be broken in a market in which oil prices were high and rising, especially if grain prices were relatively unaffected. Under those conditions, it would be much more profitable to rent property with olives, and Apollo's estates would lose value. Appeal cannot be made to vines as the decisive crop (like olives, vines need some time to begin producing once planted) because Group I estates, which entirely lacked vines, show the highest rents before 290, and Group IIA estates, which had the most vines, show the least decline in 290 B.C. Moreover, the telltale data of 280-270 B.C. deserve close attention. In that decade, rents rose while oil prices declined. This strongly suggests that the link between oil prices and rents that obtained after 270 may not have prevailed before; that is to say, the high oil prices and high rents may not be causally linked. Finally, our study of oil prices in chapter 5 has found an alternative explanation for the high and declining oil prices of 314-270 B.C.

Jacques Tréheux has offered a quite different explanation for the rent collapse of 290 B.C.^[25] Renewal of four estates in Bouphonion 314 B.C. by Delians, evidently for a 10 percent increase in rent, suggests that

renewals practiced during independence were also permitted un-

[23] See Appendix IV, p. 306, below. Glotz, Journal des Savants 11 (1913): 19–20, with REG 31 (1918): 214; Jardé, 169–70, 170 (cf. Shear, 31); Larsen, 383–84; Heichelheim, Wirt. Schw., 51, and "Sitos," cols. 857–58; Roebuck, CP 40 (1945): 159–61; Kevin Clinton, Arch. Eph. (1971): 110–11; Foxhall-Forbes, 53–55; Garnsey, 25. But see G. Reger, Classical Antiquity 12 (1993): 304–14.

[24] Garnsey, 154–62.

[25] Tréheux, "Dernières Années," 1011–22.

---

der the Amphiktyonia. Applying successive 10 percent increases to rents known for some estates in 350 B.C. (ID 104-11A8-21 [= IG II² 1638]) yields rents close to those attested for the same estates in 315 B.C. , the last year of the Amphiktyonia (IG XI 2.135). From this, Tréheux argues that renters accepted 10 percent increases at renewal because the real rental value of the properties had risen by more than 10 percent. Thus, prices must have been rising too since at least the late 340s, well before the Greek expansion to the East. The Delian rises of early independence would just continue this phenomenon.

But this view is not wholly satisfactory either. First, for most of the estates, the really impressive rises occurred, not on the cusp of independence, but between 310 and 300 B.C. , when rises of 20–40 percent were typical and 75 percent not unknown (see table 6.1). Thus it seems evident that the truly dramatic rise in estate rents was not connected with the late Amphiktyonia, but with the early years of independence. Second, if a substantial number of renters were Athenians, rents may have been set in part by comparison with costs in Athens. Third, factors other than purely economic ones may have predisposed renters to retain their estates; there may have been some prestige associated with renting land on sacred Delos. In the absence of detailed knowledge of the rent histories of 340-320 B.C. , it is impossible to sort out such considerations. Fourth, this explanation begs the question to some extent, for it still fails to present the reasons why rents should have been rising.

J. H. Kent and Michèle Brunet have taken related approaches to the problem. Kent argued that defaults by renters before 290 led the hieropoioi to try to protect Apollo's interests by proceeding against other persons who had borrowed money from the temple and failed to repay it. They seized their property, hypothecated as security, and these processes discouraged other renters. He explains the high rents themselves by suggesting that the Delians were "possibly moved as well by a false optimism engendered by their newly acquired independence."^[26] Brunet, who saw how unsatisfactory this account was, postulated instead that a rising demand for estates after 314 (more on this later), which fueled the rise in rents, drove the hieropoioi to seize insolvent debtors' property as a mechanism to satisfy demand for more estates to rent.^[27] But her view, no more than Kent's, fails to explain the connection between high rents and the failure of debtors. In an "inflationary" economy (Brunet's term), debts should

[26] Kent, 284–85; 308 (quotation), followed by Cavagnola, Istituto lombardo, rendiconti 107 (1973): 539–40.

[27] Brunet, 62–64; cf. esp. 63.

---

be easier, not harder, to pay, especially for estate owners presumably benefiting from high agricultural prices.^[28] Moreover, she offers no explanation for this expanded demand for the estates, although she does properly insist on looking for it in local circumstances, and not, like Heichelheim and his followers, in some general universal price rise.

Most important, none of these views really comes to grips with the two most puzzling aspects of the problem: why should the collapse have occurred exactly in 290, and why did different groups of estates behave differently? I would like to explore a very different kind of explanation. I argue that the Delians, "locked out" of investment in sacred properties during the Amphiktyonia, rushed to seize the opportunity once the Athenians departed. (The fact that the greatest rent rises occurred between 310 and 300 B.C. supports this view.) Rapid turnover of renters coupled with the desire to have property pushed estate values up precipitously until some defaults occurred. The Delians, realizing that the estates were overvalued, promulgated new regulations for estate rental, the hiera syngraphe (ID 503), to protect Apollo's interests. These regulations greatly decreased the desirability of the estates, and hence lowered the rents; in exchange, Apollo's interests were strongly protected, and only the wealthiest and best-connected families would be able to compete for the estates.

Rental Practice in the Later Amphiktyonic Period (350-315 B.C. ) and during Early Independence (314-290 B.C. )

Estate rents had very probably been rising on Delos since about 350 B.C. Jacques Tréheux has offered a persuasive, although not conclusive, argument that most estates were renewed for an

(a 10 percent rise in rent) in 340, 330, and 320 B.C. Even though it does not follow that the "real value" of the estates must have been substantially higher, such renewals would have established a pattern of increases at renewal that might have persuaded Delian bidders after 314 B.C. that even higher rents were reasonable, laying the groundwork for the really spectacular rises of 310-300 B.C.

Since the Athenians (with, from time to time, the help of Andrians) controlled the sanctuary in this period, it seems reasonable to ask whether the Athenian administrators favored their countrymen in renting the estates. At first sight, the evidence for the origins of renters (whether of estates or houses) does not seem to support the view that Delians had trouble renting Apollo's property. The evidence is unfortunately exigu-

[28] Kent, 302 n. 199; Brunet, 62.

---

ous;^[29] only a handful of inscriptions give ethnics for renters. In the case of the sacred houses, one document (ID 104-8B1-51) shows probably a total of seventeen Athenian renters (including one Athenian metic) against nine Delians.^[30] The six inscriptions that offer some evidence about the estates yield only four Athenian renters as opposed to nine or ten Delians and one Tenian.^[31] Delians seem to have done well against their Athenian competition.

But it would be a mistake to conclude from this evidence that there was no pent-up demand on Delos for Apollo's property. The ratio of Delian to Athenian renters seems to have changed from rental period to rental period. ID 104–21, which dates to about 346 B.C. ,^[32] reports only one Athenian renter against five Delians. But ID 104–26, which may date to 350 B.C. ,^[33] has three Athenians and two or three Delians. Tréheux has pointed out that four estates seem to have been in the hands of Delians in 314 B.C. , for they were allowed to accept a rent rise of 10 percent in return for not having to bid competitively for the estates.^[34] The rest of the estates passed into new hands. It may be that some of the other estates were in the hands of Delians who preferred to take their chances in the bidding and lost out, but it seems more reasonable to me to suppose that the renters had been Athenians, now dispossessed. The pattern of renewals by

that Tréheux reconstructs would have made the procedure seem normal to those in a position to take advantage of it, like the four Delians.^[35] There is evidence of tightening Athenian control after midcentury, and particularly

[29] The discussion of the evidence by Tréheux, "Dernières Années," 1021–22, has been superseded now that the texts of ID are available; cf. also J. Coupry in Atti del terzio congresso (Rome, 1959), 63–64. Unfortunately, the detailed treatment promised by Coupry in his Athènes et Apollon Délien: Recherches sur l'histoire de Délos, de l'époque de Pisistrate à l'époque d'Alexandre le Grand has not yet appeared.

[30] Cf. Coupry's comm. at ID, p. 64. Tréheux's conclusions ("Dernières Années," 1021) depend on the less reliable text at BCH 29 (1905): 423–24, no. 140. On the date, see J. Tréheux, EAC 5 (1976): 92 n. 56.

[31] ID 98Ab105, 104-19A7-14, 104–20 (rerental after a default, cf. Coupry, ID, p. 83, and Brunet, 26–27), 104-21bA10, 12 and bB5–21, 104-26A9-24, 104–26bis A'. Cf. Coupry, ID, pp. 87, 104.

[32] See J. Coupry, Atti, 64.

[33] Brunet, 36.

[34] Tréheux, "Dernières Années," 1011, 1022. The practice is permitted by the hiera syngraphe (IG XI 2.287A174; the apposite clauses of ID 503 have been lost, but see Kent, 270), and was evidently already allowed under the Amphiktyonia if Tréheux is correct; cf. also Kent, 260 n. 49, who suggests that the rent figures of ID 102.8–9 (ref. to BCH 8 [1884]: 313, no. 15) may reflect 10 percent renewals.

---

after the decision of the Delphic Amphiktyonia in 344 B.C. confirming the Athenians' right to control the temple.^[36] It is therefore quite possible that only four estates were in Delian hands in 314 B.C. The impact of even a few Athenian renters should not be underestimated. Apollo's patrimony included only eight or ten estates under the Amphiktyonia, and the right of Athenians to rent them reduced the number available to locals. The fierce competition among Delians after 314 B.C. strongly suggests that even a modest Athenian contingent would have restricted the locals.

There is another, even more important factor. As Jacques Coupry has emphasized, all guarantors certainly identified, whether acting for renters of estates or renters of houses, were Athenians.^[37] Since no one could rent sacred property without guarantors, all Delians without strong ties in the Athenian community were effectively debarred from Apollo's property. Only Delian "collaborators" could exercise the theoretical right of Delians to rent from the god. We know from documents like IG II² 222 the hostility that a Delian's friendliness to the Athenians could provoke among his fellow citizens. It would not be surprising if only a small minority of those Delians wealthy enough to rent property had the taste for collaboration—and the connections in Athens—necessary to indulge themselves.

In any case, there can be no doubt that the Delians moved aggressively in the fall of 314 B.C. to extirpate every remnant of Athenian control from their newly liberated island. As Tréheux has shown, they cancelled leases in force in late 314 for the

that had been let out under the Athenians in 315 B.C. and replaced them with temporary leases good only until the beginning of the next Delian year in Lenaion. Estate leases were cancelled as well and replaced with new four-year leases that ran until 310 B.C.^[38] Tréheux has even suggested that the Delians may have confiscated the Andrian as punishment for Andrian cooperation with the Athenian masters.^[39]

[36] Cf. ID 104–22, with Coupry's comments, pp. 89–90.

[37] Coupry in Atti, 64. One possible exception at ID 98 Ab105.

[38] Tréheux, "Dernières Années," 1011–12, 1028–31.

[39] Jacques Tréheux in Stemmata, 386.

---

Over the next few years, leases were renewed on a quite irregular basis, as we have seen, with four different leasing periods between 315 and 300 B.C. The irregularity of rental strongly mitigates, in my view, against the presumption that the regulations that had covered estate rentals during the Amphiktyonia continued in force after 314 with little break.^[40] Although the hieropoioi accepted renewal with an

certainly once, and possibly twice more,^[41] the abandonment of the ten-year leases of the amphiktyonic period, the relative rarity of 10 percent renewal, and the apparently ad hoc procedures followed during foreclosures for failure to pay the rent (see below) strongly suggest that the hieropoioi had abandoned Athenian administrative practice. Their willingness to do so may be related to the dramatic rises in rents after 314 and especially 310 B.C.

Competition among renters for the estates was fierce during these two decades. It is unusual for the same man to rent an estate for two consecutive lease cycles (only three or four instances out of seventy-five rental periods [fifteen estates X five rental periods each]). Table 6.1 shows how often and by how much competing renters were willing to bid each other up for control of the estates on very short-term leases.

The lessees belonged to the highest circles of Delian society. Twenty-eight out of fifty-nine different renters for the estates between 314 and 282 B.C. can be assigned with certainty or high probability to wealthy, highly placed Delian families. Among them are Lysixenos son of Aristoboulos, arkhon of 301 B.C. and renter of Lykoneion; Skymnos, another

[40] Tréheux, "Dernières Années," 1012: "les hiéropes de l'Indépendance n'ont fait, le plus souvent, en matière juridique, que reprendre à leur compte les méthodes des administrateurs athéniens et les différences qu'on observe entre les deux gestions sont moins de principe que d'application." Cf. also Tréheux, BCH 68–69 (1944–45): 294–95. I do not see how Tréheux can claim that "the hieropoioi must have followed [the amphiktyonic hiera syngraphe] up to the publication of the new text" (i.e., the hiera syngraphe in ID 503) when, as he himself admits, they violated the ten-year lease provisions of that regulation again and again after 314 B.C. (ibid., 293). On the amphiktyonic ten-year lease, see also Kent, 259–260 (probable but not certain); Tréheux, "Dernières Années," 1012 (implicitly accepted); Durrbach, BCH 29 (1905): 443, and 35 (1911): 19.

[41] See table 6.1. In two cases, however, the "renewer" is in fact a different person; in two more (IG XI 2.142.1–4) the names of the renters are not preserved (although the rents increase by 10 percent between this lease and the one represented in IG XI 2.144), and given the first two cases, there is no warrant to assume that the same persons were renters. The only solid case is that of Panormos, rented by Maisiades in 304 B.C. for 925 dr and again in 300 and later for 1,030 (IG XI 2.144A12, 149.5–6, 147A15–17). On the practice, see Maurice Lacroix, BCH 56 (1932): 373; J. H. Kent, BCH 63 (1939): 233; and Tréheux, BCH 68–69 (1944–45): 285.

---

renter of Lykoneion, who is probably the homonymous hieropoios of 298 B.C. ; Empedokles, renter of Nikou Khoros, who guaranteed a state loan; and Amnos son of Dexikrates, who moved two decrees in the assembly and rented Skitoneia. These were the people who controlled Delian political life; for them, as a traditional Greek aristocracy, investment in land was an important social desideratum. As soon as the opportunity arose, they rushed to capture sacred leases, and the competition was apparently fierce.^[42]

This competition must at first have appealed to the hieropoioi charged with the administration of the sacred estates. They could offer quite short leases—perhaps as little as two years—permitting rapid turnover in renters, and usually guaranteeing increasing revenues for Apollo. Already by the second lease period, however, some potential difficulties had begun to emerge. Sometime during that lease period, two renters failed to fulfill their obligations and were dispossessed (IG XI 2.142.5–12). In neither case were the hieropoioi able to exact the full rent due by sale of crops and farm animals, or by exaction from guarantors; Apollo took a loss of at least 645 dr. Additional payments, called

in IG XI 2.142.2–4, may indicate difficulties with other estates; the year was unusual enough to be remembered almost twenty years later as .^[43] In 304 B.C. , there was another default, but because the crucial figures are missing, it is not possible to know whether the hieropoioi succeeded in exacting the full amount from the renter's guarantors.^[44] Four years later, yet another renter defaulted; the hieropoioi record his guarantor's partial payment of the back rent (IG XI 2.147A15–17). The record shows four defaults in ten years.

It is not difficult to imagine the problems these losses posed for the Delian community. On the one hand, Apollo's interests wanted protection. The god looked for steady, reliable income from his estates, and that meant above all renters and guarantors wealthy enough to cover any potential losses. On the other hand, the Delian aristocrats who rented the estates, and especially any belonging to levels of Delian society under the very top, had social and economic interests in the prevailing system, which permit-

[42] Kent, 330, no. 145 (IG XI 2.145.46); Kent, 334, no. 211, Vial, 263, Stemma XXI; Kent, 326, no. 83, Vial, 334; Vial, 133, with n. 31. For full details of the renters, see Kent, 320–38; Vial, passim; and Bruno Cavagnola, Istituto lombardo, rendiconti, Class. di lett. e sc. mor. e stor. 106 (1972): 51–115, esp. 90–110. Cf. also Robin Osborne, Chiron 18 (1988): 299–300. On the social position and attitudes of upper-class Delians, see the detailed and definitive study by Vial, 253–74, 283–306, 317–56.

[43] IG XI 2.156A14; Tréheux, BCH 68–69 (1944–45): 284–85.

[44] IG XI 2.144B13, 78–81. The estate has evidently been rerented at II. A13 and B78–81; cf. J. H. Kent, BCH 38 (1939): 235.

---

ted many different people to profit from Apollo's holdings, albeit at rather high risk.

The Impact of the Hiera Syngraphe (ID 503)

The hiera syngraphe, first put into effect with the leasing period of 299-290 B.C. ,^[45] seems to have been designed to solve these problems. The text preserved sets out in great detail regulations to govern the rental of Delian estates. Among the provisions are stricter requirements for guarantors (who now must be renewed annually) and detailed regulations covering payment of rent and default. As Kent suggested many years ago, these regulations seem clearly intended to prevent further serious defaults.^[46] In particular, two strong, broad clauses subjected all the belongings of the renters and their guarantors to seizure:

(crops, plow oxen, livestock, and slaves) [the hieropoioi ] and ,and .^[47] Had the hiera syngraphe been in effect in the 300s, this clause would have permitted much more comprehensive action against the defaulters of IG XI 2.142.5–12, whose personal goods were not expropriated, even though Apollo failed to get full restitution after the seizure of Hermadas's barley crop and Arkhandros's barley crop and plow team. Under the hiera syngraphe, a renter and his guarantors who failed to pay the full rent, or violated certain other regulations, might theoretically find not just the investment in the estate—crops, oxen, livestock, and slaves—but their entire personal fortunes in jeopardy. The threat alone might have made rental of estates seem riskier after 300 B.C. But did the hieropoioi ever in fact proceed with such vigor against defaulters?

[45] On the date, see Appendix I.

[46] Pierre Roussel, Délos colonie athénienne (Paris, 1916), 73 (on the attitude of the Athenian administrators); Kent, 279–80, 285; Tréheux, BCH 68–69 (1944–45): 290–92, 295 (pointing out however that not all provisions were strengthened, only those covering default); Kent, 267–89; Marie-Françoise Baslez, REG 89 (1976): 347–48; Dieter Hennig, Chiron 13 (1983): 442–43 n. 71. Brunet, 62; contra, Erich Ziebarth, Hermes 61 (1926): 96–97, whose views Tréheux answers, however. Cf. also Felix Durrbach, BCH 35 (1911): 25–29, and René Vallois, BCH 55 (1931): 290–91.

[47] ID 503.34–36, 46–48; cf. Kent, 282.

---

A good deal of circumstantial evidence strongly hints that they did.^[48] I argue that the efforts of the hieropoioi to rosecute defaulters under these new and stronger regulations had two important consequences: (a) the possessions of Apollo were increased by seizure of land and houses from defaulting renters, which (b) helped considerably to reduce the appeal of the estates, and hence the rents to which renters were willing to commit.

Let us consider three cases.

(1) In 297 B.C. , one Autosthenes was renting Dionysios for 1,372 dr. His guarantor was Kleokritos son of Hermon (IG XI 2.149.6–7, 11–12), whose prominent and wealthy family counted among its members a hieropoios of 297 and a khoregos in 255 B.C. The family had extensive interests in the sacred estates, three members having rented four different estates at different times.^[49]

In 314 B.C. , Kleokritos's father, Hermon son of Kleokritos, paid interest of 150 dr on behalf of one Sosimakhos (IG XI 2.135.26–27). Hypothecated land or a house or both must lie behind this payment. Payments "on behalf of someone" were usually made either by guarantors or heirs (the guarantors themselves often being relatives with an interest in the hypothecated property) or by outsiders who bought the hypothecated property and assumed the debt that accompanied it.^[50] Thus it is fairly certain that Sosimakhos's property had come into the hands of Hermon's family by 314 B.C.

If Autosthenes had defaulted at a time during his rental of Dionysios when Kleokritos was his guarantor, Kleokritos's property would have been "subject to the god" under the rules of the hiera syngraphe. If Autosthenes' debt could not be collected from his livestock and slaves, the hieropoioi would have proceeded against his real goods and those of his guarantors.

[48] Vial, 224, rejects the possibility: "il est peu vraisemblable que les hiéropes aient jamais saisi les biens personnels du fermier et de ses garants" and at n. 137 cites the hieropoioi's failure to seize the personal property of Hermadas and Arkhandros in 305 B.C. or slightly earlier. But the hiera syngraphe was not in force before 300 B.C. , and Vial's parallels—failure to act against defaulting house renters, tax-farmers, or borrowers—founder on our ignorance of the conditions that governed these transactions. (If ID 499 is a regulation governing loans, it remains unfortunately too fragmentary to be very informative.) The evidence outlined below, if correctly interpreted, outweighs these theoretical considerations.

[49] See Vial, 322, Stemma XXXI, for full details and references.

[50] For inheritance, see, e.g., IG XI 2.135.19–20 and 142.14; 199A11 and many other entries, with Vial, 82–83, Stemma XII, 372, no. 124; for the case of Apollodoros of Kyzikos, who bought a hypothecated garden (IG XI 2.142.14–15), see G. Reger, GRBS 32 (1991): 229–37; another example at IG XI 2.287A15 with Vial, 295, Stemma XXV.

---

In this way, Sosimakhos's land, now in Kleokritos's family's possession, might have fallen into the hands of Apollo, becoming the sacred estate Sosimakheia, known first from the rental period of 289-280 B.C.^[51]

There is more. It would be strange if the hieropoioi had foreclosed on Kleokritos without also acting against the renter himself. Among the renters of the sacred houses appears the name of Autosthenes, who rented "the house that belonged to the children of Aristoboulos" (

) in 279 and 278 (IG XI 2.161A18–19, 162A18). Like many other houses, this one is not attested during the Amphiktyonia, appearing first only in the 280s (IG XI 2.156A1, 158A23). It was not unknown for former owners whose houses had, for whatever reason, gone over to Apollo, to continue to occupy and rent them.^[52] If Autosthenes had done the same, then this house could have been his contribution to his debt to Apollo.

This reasoning makes of Autosthenes a son of Aristoboulos. In fact, an Aristoboulos, who is twice specified as Aristoboulos son of Lysixenos,^[53] rented the same house—undoubtedly originally his own—from 269 to 246 B.C. This man, who probably also leased one of the

in 262 B.C. and failed to pay his rent,^[54] was a prominent personage with important temple connections: he was priest of Asklepios in 279 and may well have

[52] The house of Antigonos, IG XI 2.158A20; the house of Arkeon, 158A16–17, 161A23, 162A17, 199B94; perhaps the house where Ephesos had his shop, 161A13–14, although it was rented by Aristolokhos in the 280s (157A14, 158A21): perhaps Aristolokhos was Ephesos's owner and manumitted him sometime between 282 and 279 B.C. ?

[53] IG XI 2.226A19, 287A36; further 203A27, 204.34, 224A20–21, ID 290.23.

---

been the Aristoboulos who served as an epimeletes in 274 B.C. He rented Soloe in 289–280 B.C.^[55]

Finally, it was typical for guarantors to have some connection with renters. The very first attested renter of the house of Aristoboulos was one Diakritos (IG XI 2.156A1, 158A23). The name recurs, and prominently: it belongs to Kleokritos's brother, who was hieropoios in 297 B.C. , the very year Kleokritos was named by Autosthenes as his guarantor (IG XI 2.148.77, perhaps for the first time).

None of this is proof, of course; what we most sorely lack is direct evidence of a default by Autosthenes. The circumstantial evidence, however, is strong, and a default and procedure by the hieropoioi (obviously not the board with Diakritos on it!) against Autosthenes and Kleokritos would explain neatly the appearance of two new properties among the possessions of Apollo.

(2) The renter of Nikou Khoros in 297 B.C. was a man named Sosilos (IG XI 2.149.5). Like Kleokritos and Autosthenes, he belonged to an important family.^[56] Through a close relative (perhaps his cousin?) and his son Gorgias, the family bought up, between about 282-279 and 274-250 B.C. , a piece of property at Passiros (or Passiron) that had belonged to the family of Eurymanthes.^[57] It is tempting to connect this family, and Sosilos in particular, with a series of houses owned by Apollo and called collectively

. They first appear in the 280s, although one had collapsed and was not rented until 252.^[58] If Sosilos suffered a default in the 290s on Nikou Khoros, his family may have lost their house(s), and the purchase of Eurymanthes' family's land at Passiros may have been designed, in part, to make up for the loss.

(3) Phytalia is another estate that first appears in the accounts during the rental period of 289-280 B.C.^[59] On the basis of a restoration in an account of the Amphiktyonia, some have thought that Phytalia was one of the estates confiscated in 375 B.C. and returned to its owner's family in

[55] IG XI 2.161D3–12; 199A82–83 with Vial, 165 n. 9; 157A3, 158A12–13.

[56] Vial, 136–37, Stemma XVI, cf. also 326–28.

[57] Vial, 326–28, with Stemma XXXII, p. 327. Sosilos was the son of Mnesalkos, a name that recurs in the family of Stemma XVI. If this Mnesalkos was Telesarkhides' brother, then Sosilos and Mnesalkos I would be cousins (cf. Vial, 136, Stemma XVI).

[59] IG XI 2.158A14, rented by Philtes.

---

314 B.C. Tréheux and Brunet have expressed doubt.^[60] Kent thought that, like Sosimakheia, it had come into Apollo's hands as a result of unpaid debt. As we shall see, this view cannot be supported, but the evidence Kent adduced does point to the solution. In 250 B.C. , Diaktorides son of Theorylos borrowed 400 dr from Apollo

(IG XI 2.287A129–31).^[61] Kent's view, that this entry shows that Phytalia was seized for debt, is insupportable: the property called Phytalia used to belong to Pherekleides, but nothing is said about how he lost it. Indeed, Claude Vial thinks "Phytalia" has nothing to do with Pherekleides at all. She understands the relative clause to refer not to the same as but to the : "on the hypothecation of land [] that borders land [] that used to belong to Pherekleides, and that is called Phytalia."^[62] This understanding is forced against what seems to me the plain meaning of the whole sentence: Diaktorides borrows 400 dr "on the hypothecation of [a] the land on which borders the land that used to belong to Pherekleides and that is called Phytalia, and on [b] all other things that exist for Diaktorides, and [c] on his guarantors Kallisthenes son of Theorylos, Antigonos son of Didymos."^[63] Phytalia, which once belonged to Pherekleides, has been used to delimit a neighboring property in private hands that has been hypothecated for a loan.

If this is right, it is easy to find circumstances under which Pherekleides

[61] Kent, 286, with n. 149, where he restores the same expression at ID 287bis 20. Vial, 326 n. 56, mentions this entry without Kent's restoration.

[62] Vial, 326, followed by Brunet, 58–59. Bogaert, 152 n. 112, cited by Vial and Brunet, writes only, "Phytalia a en effet appartenu à un certain Phérécleidès . . . que M. Kent identifie avec le débiteur dont nous avons parlé supra p. 150. Mais il faut remarquer que ce domaine était devenu propriété du dieu avant 290 et que la dette de Phérécleidès existait encore en 250. Il ne peut donc s'agir de la même créance." His refutation of Kent proves only that Pherekleides' own personal debt cannot have been the cause of the loss of Phytalia, not that the estate cannot have come into Apollo's hands by the route I lay out below.

---

could have lost his property. In 297 B.C. , his two brothers Proxenos and Khares independently rented Leimon and Phoinikes. The default of either one could have led to the seizure of Pherekleides' land.^[64]

Two other estates, Korakia and Akra Delos, also appear for the first time, and Epistheneia reappears,^[65] in this decade. It may fairly be suspected that Apollo obtained them by the same route.

If these seizures are real, they would certainly have made a severe impression on potential renters. The threatened—and sometimes all too real—loss of personal property must have frightened both renters and their guarantors. The greatly lengthened leases of ten years, at least double the terms that had prevailed in 314-300 B.C. , added more uncertainty: who could be confident that a decade would not see at least one year when unexpected losses would make meeting payments difficult, if not impossible? After 290 B.C. , most of the defaults we hear of resulted from the failure of renters to renew guarantors; given the potential losses for guarantors in a bad year and the absence of any benefits during good ones, the problem is not surprising.

Unfortunately, the seizures I have postulated to explain Apollo's acquisition of new estates in the 290s are only hypothetical. Is there any positive evidence for such procedures during 299-290 B.C. ? The very poor state of the inscriptions for that decade renders definite conclusions difficult, but luckily a few indications do survive.

The most important is IG XI 2.152. In the absence of an arkhon's name or other definite indicators, it must be dated on style of writing, internal grounds, and other criteria. In style it fits well with other documents of the 290s; the few names point to the 290s or the 280s.^[66] Of these, Diaitos may be identical with Diaitos son of Apollodoros, known from ID 502A29 of 297 B.C. and other documents of the next twenty years.^[67] Diaitos is recorded here as a renter of an estate. Since no such renter is known from any period, he must have rented an estate for only part of a decennium: either he took over from a defaulter after 297 or he rented an estate in 290 and defaulted before 282 B.C. (I leave aside other imaginable, but more complicated, scenarios.) Of the two possibilities, I prefer the first. The few

[64] IG XI 2.149.3–4; Vial, 295, Stemma XXV. Starting in 289, Pherekleides himself rented Leimon, which Proxenos had held in 299-290.

[66] Theophile Homolle, Les Archives de I'intendance sacrée à Délos (Paris, 1887), 119, no. XI.

[67] IG XI 2.161A41, D82, 83; 203D77, 78; perhaps 199C85.

---

words of 152A7, [

], more closely resemble the language the hieropoioi use of rerenters than that for defaulters.^[68] This would put 152 in the 290s, but this argument is very tenuous.

The content of the document is arresting. Despite the very fragmentary state of the text, it clearly records judicial proceedings against renters: it mentions the boule (

,  A3); a debt and a law court ([], A5–6); perhaps interest of 10 percent, but perhaps better a reference to 10 percent renewals ( in IG, which may be better restored as , A8); at least three estates, including Kharoneia (, A9); the renters (, A10; cf. IG XI 2.147A18); the deprivation of something (, A12); rental (, A13–14); at least 1,400 dr (, A14–15); a total figure (, A15); and a payment by a someone in the first person singular to the hieropoioi (, A16).

In my view, this text records the trial and conviction of renters who violated the rental contract, payment of fines and/or back rent, and the seizure of property from those who could not or would not pay. Restoration at A12 as

, someone "will deprive [the renter?] of his [land]," is almost unavoidable. The last phrase may belong to an oath the condemned were forced to swear about payments.

Unfortunately, the absence of the names of all but one of the estates involved in IG XI 2.152 stymies further inquiry. For Kharoneia we have only hints. The figure at A14–15, 1,400 dr, might refer to the rent: one Xenomedes was paying 1,450 dr/yr for it in 297 B.C. (IG XI 2.149.7–8). His name recurs twice in contemporaneous documents. A Xenomedes son of Apatourios served on the commission of Eleven in 297 B.C. overseeing certain contract work done for the temple. The name also occurs as a purchaser, with two guarantors, in IG XI 2.153.^[69]

IG XI 2.152 does not stand alone. Another account, probably of almost the same date, attests to a similar disaster, in which at least three renters defaulted (IG XI 2.153.21–27). The guarantors are explicitly included in the proceedings; we have the names of two, Aristodikos and Hypselos (ll. 22, 23), unfortunately not otherwise attested in connection with the

[69] ID 502A27; see Vial, 116–19, on the board; IG XI 2.153.17. It is not possible to trace the guarantors, Kallimos son of Patrokles and Dionysodoros son of Lysileos, farther, although Kallimos was perhaps the synonymous arkhon of 268 B.C. (IG XI 2.110.17–18). Neither Vial nor Brunet discusses IG XI 2.152.

---

estates.^[70] The miserable fragment IG XI 2.151 may preserve a third reference to this business at 1.6 in the word

.^[71]

These cases are not unique.^[72] In 247 B.C. , two renters were tried and convicted, one by unanimous vote of the jury. The one defendant had failed to pay rent,

, and the other had destroyed something, ; Vial thinks of either buildings or trees.^[73] Unfortunately, it is not possible to recover the name of either defendant; one is lost completely, and the other, , is not among the renters for the decennial 249-240 B.C. , fully preserved at IG XI 2.287A142–180; he probably took over from another defaulter in 249 or 248.^[74] In another similar case in 206 B.C. , probably three renters were fined one and a half times the rent owed (). Here again, it seems likely to me that these men did not have sufficient property to cover the full rent owed.^[75] These examples show indisputably that the hieropoioi had both the authority and the will to proceed by law against renters who had violated the terms of the hiera syngraphe.^[76]

In the most generally accepted view, however, the new estates of the early third century came into Apollo's hands through a very different process: they were seized from borrowers who had hypothecated them to Apollo for loans they failed to repay. The evidence to support this view comes from IG XI 2.135, where one Tharsynon son of Hierognotos paid 200 dr interest

(ll. 22–23), clearly private property. The inference that this land became the estate Epistheneia after

[70] A Hypselos appears as a contractor in IG XI 2.145.18.

[72] I do not accept Ziebarth's very speculative restorations for IG XI 2.225b7 and 199D32–35 (ibid.).

[73] ID 291d35, 33, with Vial, 156, 230; cf. Brunet, 71.

[75] ID 369A40–41, cf. Brunet, 71. For what may be a similar case under the Amphiktyonia, see ID 104–19A.

[76] Although the surviving clauses of the hiera syngraphe contain no provisions about the maintenance of capital goods on the estates, numerous parallels make it virtually certain that there was such a stipulation: IG II 1241.30–33, 2492.14–17, 2494.11–16, 2496.15–17 (renter to repair structures as necessary), 2499.14–18 (Athens), IG XII 5.568.11–14 (Poiessa on Keos), IG XII 7.62.8–13 (Arkesine on Amorgos; care enjoined rather than cutting prohibited); cf. Kent, 272; Robin Osborne, Classical Landscape with Figures (London, 1987), 42–43. Vial, 230, claims that the hieropoioi carried out repairs on the estates, but can cite no evidence (cf. ibid., 144).

---

default on the loan is not hard to draw. The interest paid

(ll. 26–27), which I have already discussed, is supposed to provide another example.^[77]

There is one insurmountable objection to this view: never in the entire history of the island did the hieropoioi proceed against private debtors who had defaulted on their payments. The leniency with which Apollo treated slackers extended even to conferring new loans on defaulters; no one who failed to pay his annual interest payments was ever, to our knowledge, debarred from further traffic with the temple, whether as contractor or renter or official; and certainly, no property can be shown ever to have come into Apollo's possession because of action against a defaulting borrower.^[78] Indeed, the owner of the property Sosimakhos had hypothecated may have himself benefited from this leniency. ID 104-8A15, which can be dated only to 360-330 B.C. , carries a payment

. If this refers to the same loan recorded in 313 B.C. , Tharsynon paid off ten years' back debt (or 150 dr), covering presumably 322-313 B.C. Yet during those ten years, while unpaid interest mounted up, neither the Athenian administrators nor their successors acted to seize the security.

To see the seizures of Akra Delos, Epistheneia, Korakia, Phytalia, and Sosimakheia.^[79] as a consequence of the hiera syngraphe, high rents for the estates, and subsequent defaults permits a very satisfying reconstruction of the events of the decade from 300 to 290 B.C. The defaults of the 300s must have frustrated the hieropoioi. Their inability to seize property prevented recovery of unpaid rent, and the relatively low risks to renters, who gambled only the investment made directly in the estates, did nothing to discourage the ever-increasing competition for the estates. By 301 or 300, the Delians at last extended the power of the hieropoioi to move against defaulters by the publication of the stricter regulations of the hiera syngra-

[77] Jardé, 147 n. 1 (with his minority view that the Epistheneia of independence had no relation to the Epistheneia confiscated in 375 B.C. ), cf. Heichelheim, Wirt. Sch., 134–35 n. 1, G. Glotz, REG 45 (1932): 243, Larsen, 405; J. Tréheux, "Dernières Années," 1016 n. 2; Kent, 256–57, 286 n. 149, with however the notion that Epistheneia and Kerameion were confiscated in 375 B.C. and never returned but registered among the houses (impossible for Epistheneia: see Appendix I; Kent explains IG XI 2.135.22–23, unconvincingly, as perhaps "interest on a loan that was raised on other security in order to avoid forfeiture of the lease of Epistheneia" as a house [Kent, 257 n. 38]); Bogaert, 152 n. 112; Vial, 224 n. 139; Brunet, 61–63.

[78] Bogaert, 138–65, esp. 151–52: "Pour aucun immeuble acquis depuis l'indépendance, nous ne possédons un document qui permette d'affirmer qu'il aurait été confisqué pour dettes"; Vial, 374; Brunet, 60–61.

[79] Always excepting Kerameion, which had evidently remained in Apollo's possession ever since 375 B.C.

---

phe. At first, having no experience with the new rules, renters proceeded as before: they bid high for estates in 300 B.C. , driven, as I have argued, as much by social as economic motives. But now the situation had changed. When the first inevitable defaults ensued, the hieropoioi took the renters to court, secured their condemnation, and seized real estate belonging to them and to their guarantors. A few such cases—recorded in the laconic but allusive inscriptions IG XI 2.151–53—chilled demand: fear of loss of property in the case of default dampened prices, since default would be less likely if rent was lower, and scared off potential guarantors, whose interests in the whole business were marginal anyway. The paradoxical result was to make the estates appealing only to the richest members of Delian society: men whose personal wealth reduced the likelihood of default, and whose connections with other wealthy men (especially their immediate relatives) eased the problem of finding guarantors.^[80]

We can read the impact of these events in the rents of the estates. As figure 6.1 and table 6.1 show, between 297 and 290 B.C. rents plummeted by an average of more than 43 percent. For the rest of the third century, rents tended to be stable; moreover, the same renters often rented the same estates for more than one rental period or passed estates on to relatives. Over the following 110 years (to 180 B.C. ), there were eleven renewals by families and twenty-six by individuals out of two hundred rental cycles (twenty estates and ten rental periods each), a rerental rate of 18.5 percent, as opposed to 5.3–4.0 percent before 290 B.C.^[81]

Defaults do not seem to have become any rarer, as the record shows thirty-one certain or probable instances from 290 to 175 B.C. , but their character changes: in the vast majority of cases where details are preserved—and that accounts for most of the defaults—the renter defaulted because of failure to renew his guarantors. In every case the hieropoioi found some-

[80] I think this account of the decline in rents far more plausible than Brunet's notion (61–64) that the new estates just exactly matched unsatisfied demand, and so depressed rents. The coincidence that six new estates exactly balanced demand seems incredible (there were then only six or seven unsatisfied bidders in 300?). Moreover, if this were so, one would expect the renters of 300 B.C. to continue renting in 290 (since there were now plenty of estates to go around), but in fact only one renter of the 299–290 period, Aristeides who held Hippodromos, continued in 289–280 B.C. Moreover, would not the expelled owners of confiscated estates, now deprived of their property, enter the bidding for estates (like exproprietors of the sacred houses) and thus increase demand, counteracting the expanded supply?

[81] See Appendix IV, pp. 309–38, below. I do not see how Robin Osborne, Chiron 18 (1988): 300, arrives at only eight cases of renewal by individuals and three by families.

---

one else to assume the lease, although sometimes for a reduced rent. In these cases the former renters (or their guarantors) were held responsible for the difference. The regulations seem to have protected Apollo's financial interests well.^[82]

This stability had advantages for both Apollo and the Delian aristocracy. By eliminating the fierce and expensive competition that had prevailed in the two decades after independence, it minimized the impact of defaults and losses for Apollo. Renters were now able to hold estates over longer periods, and the ten-year lease, which virtually guaranteed that renters would face bad years during its term, would doubtless have discouraged speculators from bidding too high. The losers in the game were probably the less wealthy members of the Delian upper class who had neither wealthy relatives to stand surety nor the resources to cover rents in years when crops failed.^[83]

Aside from its inherent probability, this reconstruction of events has the advantage of looking to local circumstances for an explanation of the boom and fall in prices of land, and is independent of any (possibly specious) account of prices for agricultural commodities on Delos or elsewhere in the Aegean. Local prices for agricultural commodities were, however, another issue: they mattered crucially to renters, as we saw above.^[84]

[82] Defaults are recorded or inferred for Epistheneia in 288–283, 206?, 178–175 B.C. (IG XI 2.156B7–15, ID 369A41, 467.1–3); Hippodromos in 262, 189 (204.8 [cf.290.9–11], 403.51–53); Kerameion in 257 (226A36); Leimon in 257, ca. 190 (226A34–35, 406B80–83); Lykoneion in 218, 209–200 (354.35, 356bis B22); Phytalia in 207, 177 (368.26–27 [cf. 371A26, 372A18], 452.24–25); Soloe-Korakia in 177 (452.31–32, but there are problems: see Appendix IV, p. 321 below); Sosimakheia in 288–283 (156B16–20); Khareteia in 257, 250 (224A14 [cf. 226A30, 225a8–9], 287A30, 139–142); Kharoneia in 250, 189 (287A29–30, 138–39, 403.48–51); Dionysios in 207, 206, 192 (366A104–5, 369A40); Nikou Khoros in 279, 210, 178 (161A9, C116–20, 356bis A12, 445.16–24); Panormos in 279, 207, 192, 177 (161A9, C111–15, 366A105–6, 399A79–80, 452.20–21 [cf. 456A18, 440B17–20]); Porthmos in 274, 207 (199A3–4, 14, 366A102–3); Rhamnoi in 250 (287A25, 136–37); Skitoneia in 250 (287A26, 137–38, D27–28). I have not distinguished certain defaults from probable ones or counted defaults on estates whose rents are not known.

[83] Cf. Vial, 331, for a partial account of the control of the estates by individual families over more than one rental period after 290 B.C.

[84] This view does not really conflict with that of Robin Osborne, Chiron 18 (1988): 302–4. Osborne is analyzing data from the later third and second centuries, not from 314–290 B.C. Moreover, commercial attitudes toward farming need not preclude a social value to estate rental. I find Osborne's conclusion rather too categorical. That, however, there may have been substantial changes in attitudes in the later third century seems very likely to me.

---

The Differing Responses of Group I and II Estates

The threat that the hiera syngraphe delivered to renters and their guarantors was the confiscation of personal property. I have concentrated on real property, houses and land, because it was the seizure of such property on a grand scale in the 290s that had such a devastating effect on the interest of potential renters. But the syngraphe envisions the confiscation of other kinds of property as well:

(ID 503.33–34).^[85] Crops, plow oxen, livestock, and slaves: these "movables" represented a renter's personal investment in his estate.

Now, although the rents all four groups of estates (I, IIA–C) fell drastically in 290 B.C. , the degree of change varied considerably from group to group. Group I estates responded most strongly, with a change of about 105 units. The three subgroups of Group II range from a mere 27 for IIA to 62 for IIB and 67 for IIC (tables 6.3, 6.5, 6.6, 6.7, "coefficient"). These results reconfirm the contention that Group I belongs apart from the others, Group IIA weathered the change with only minimal loss of appeal, and Groups IIB and IIC responded very similarly: much more strongly than Group IIA, but much less vigorously than Group I.

Not only did Group II estates respond less vigorously to the hiera syngraphe, but their rents had also risen much less before 290 B.C. than the rents for Group I estates. Indeed, the more vines an estate had, the less its rent rose during 314–300 B.C. The effect is most apparent for Group IIA estates, whose high average rent in the period is only 143. Groups IIB and IIC lie close together, as we might expect, but even here there is a noticeable difference; Group IIC average rents are invariably higher than Group IIB rents (cf. fig. 6.1, table 6.2). The sum effect of vines, then, was to dampen the swings in estate rents. In my view, two factors may be at work: the stability of the crop, and hence the income, that the vines offered; and the relatively low capital investment required of renters of estates with large vineyards.

Like olives, established vineyards can be expected to yield an approximately consistent harvest—given, of course, interannual variation owing to changes in rainfall, incidence of disease, and other factors—over the long run. As a result, renters could easily estimate the approximate income from these estates, assuming that wine prices were relatively stable. Expec-

[85] Cf. Kent, 279–80, 282. Robin Osborne, Chiron 18 (1988): 301–2, misses the significance of this clause in his discussion of the "pastoral bias" among renters.

---

tations of profit would not have varied much before or after 290 B.C. , and so the estates with vines would have been much less sensitive to the impact of the new rental regulations.

Capital investment in the production of wine would likewise have been relatively small. The vines were already there; they needed only tending, including occasional replacement of senescent stock. Cato the Elder's agricultural manual, which was written around 160 B.C. and reflected current Hellenistic thinking about farming techniques, recommends sixteen slaves for a vineyard of 100 iugera.^[86] This figure may be a bit high in comparison to later writers, who reckon one slave per 7–10 iugera, but for our purposes it can stand, implying one slave per 6.25 iugera of vineyard.^[87] Columella, who admittedly wrote at a period much later than ours, and for a specifically Italian audience,^[88] gives simple formulae for calculating plantings. He recommends placing vines at the intersections of lines laid out in the fields from 3 X 5 to 10 X 10 Roman feet. His closest planting (3 X 5) allows 2,025 vines per iugerum, his most generous (10 X 10) only 325 vines per iugerum (Rust. 5.3.1–8). If we assume that the Delian estates' vines were laid out with plenty of space to allow for the care required by the dry climate and for the cultivation of grain between rows,^[89] then for purposes of illustration, and without any claim to precision, we can estimate the manning requirements at Columella's moderate (5 X 5) and most generous (10 X 10) planting ratios (table 6.13). The latter figure corresponds roughly with Kent's own estimate of 10 m² per vine derived from observation of modern vineyards on Mykonos and Syros.^[90] Even if these estimates are low by a factor of two (which I do not regard as likely), the

[86] Cato De agri cult. 11.1. On Cato De agri cultura, see Alan E. Astin, Cato the Censor (Oxford, 1978), 189–203 (ignoring, however, the technical issues); and Silvano Boscherini, Lingua e scienza greca nel "De agri cultura" di Catone (Rome, 1970), esp. 23–91. See also Hanson in Agriculture in Ancient Greece, 161–66, on the use of slaves in viticulture.

[87] Richard Duncan-Jones, The Economy of the Roman Empire: Quantitative Studies (Cambridge, 1982), 39, 327–33; Andrea Carandini, Opus 2 (1983) 177–204. The criticism of Carandini in M. I. Finley, The Ancient Economy (Berkeley, 1985), 180–81, is not relevant here.

[88] See Carandini, Opus 2 (1983): 186–87.

[89] See Theophr. De caus. pl. 3.11–12, 10.3 (Theophrastos recommends barley). Ox sheds and other equipment related to grain production are attested for almost every estate with vines; see Kent, 299–301. Marie-Claire Amouretti in Agriculture in Ancient Greece, 83 n. 17 (planting ratios).

[90] Kent, 291–92, n. 173. For vineyards at Kydonia on Krete, cf. IC, II, Kydonia 1:2 plots of 2 plethra, 6 of 4 plethra, 1 of 10 plethra, 1 of 20 plethra, and 1 of 22 plethra. I have not seen A. Chaniotis, "Vinum creticum excellens: Zum Weinhandel Kretas," MBAH 7.1 (1988): 62–89.

---

labor requirements for the vineyards are strikingly modest. Only if vines in Group IIA estates were planted at Columella's greatest ratio would vine tending absorb the full-time labor of even one slave. Since the vines themselves and the capital equipment necessary to care for them were already present on the estates, a renter's initial investment would be confined to the cost of a single slave, whose labor might well be available much of the time for other tasks, and to a handful of other equipment, like stakes.

Group I estates present the extreme case. Entirely without vines, they offered none of the security to the renter of even Group IIC estates. Virtually all the income a renter could expect from them came from money he invested, whether as barley seed, plow oxen, and labor, or as sheep, goats, or pigs.^[91] Under the terms of the hiera syngraphe, all of this investment, along with the renter and his guarantors' personal property, was subject to seizure and loss in the event of default. Stripped of the protection afforded by the heady years of early independence, renters fled from these estates, which, it must be admitted, had a history of defaults.

The high rents and collapse of early independence can thus be explained by appeal entirely to local conditions. Social factors, always important in antiquity, take a place beside economic ones in elucidating these events. Moreover, we can now offer a very satisfactory account of why the high prices and collapse occurred when they did.

Wine Prices and Rents for Group II Estates

As we have seen above, estates of Group II all show a drop in rents after about 220 B.C. that was directly proportional to the number of vines on

---

them. As I have noted on a number of occasions, Kent attempts to explain this decline by linking it to a drop in the price of wine:

The pronounced and prolonged drop in the rentals of the vineyard estates shows that after 220 B.C. the vines yielded little or no profit, but were able to bring in merely enough revenue to pay for the expenses of maintaining them . . . yet there is evidence to show that there was no deliberate destruction of vines. . . . Since the loss of revenue by the vineyard estates was not due to loss of vines, we are obliged to conclude that the decrease was caused by a sharp drop in the value of wine.^[92]

There is certainly a great deal of truth in this view. Our results have confirmed the declining rents for "vineyard estates" after 220 B.C. and clarified the differences in rent histories arising from differing numbers of vines, a very important result. There can be no doubt that, generally speaking, estates with vines became less desirable by the last quarter of the third century, and indeed less desirable in proportion to the number of vines each had. It is far from clear, however, that this situation was a consequence of "a sharp drop in the value of wine." In fact, the evidence available for wine prices on Delos cannot be said to show any such drop at all.

For the second century, two sources offer wine prices: the accounts of the festivals of Posideia and Eileithyaia. Posideian accounts preserve three prices for Knidian wine, two for Koan, and one unidentified variety. The wines were purchased by the

, which J. A. O. Larsen identifies with the Egyptian measure of the same name of eight khoes capacity.^[93] Prices must therefore be multiplied by 1.5 to obtain the equivalent of the standard Delian metretes of 12 khoes. The results, set out in table 6.14, prove clearly that Knidian wine was more expensive than Koan by one and a half to two times. To sustain Kent's view, however, it is necessary to show, not a differential between varieties of wine (as perfectly familiar in antiquity as today),^[94] but that there was an absolute drop in price over time that depressed the price of wine produced locally on Delos and its neighbors.

For this we need a series of wine prices over time, and it does not exist. The only earlier wine price preserved from Delos—a healthy 11 dr/met —dates to 296 B.C. No variety is indicated. Gustave Glotz restored two more prices at about the same level (11 dr 4.5 ob, 10 dr 3 ob) for 304 and 279 B.C. ,

[92] Kent, 311.

[93] Larsen, 392, 394.

[94] For example, Athen. 25f–34e, esp. 28d–f on Khian, Thasian, and Lesbian, 30b–e on Ikarian, 30f on Naxian and Skiathian.

---

but as Larsen remarks they are nothing better than "plausible guesses."^[95] Even if a decline is assumed, these three prices are separated by over a century from those for the Posideia. It is impossible to know when the decline occurred. The example of oil might suggest putting it earlier rather than later: say by 270 at the latest; but this is exactly the period when Group II estates show a recovery in rents (see fig. 6.1 and table 6.2). In fact, nothing assures Glotz's two prices, and nothing justifies even the assumption that the wine bought in 296 B.C. was at all comparable to the Knidian and Koan vintages favored 100 years later at the Posideia.

Wine was also bought for the Eileithyaia.^[96] As Larsen observes, no measure is recorded,^[97] but the designation

may mean that a metretes was bought. Table 6.14 is constructed on this assumption. Philippe Bruneau and Philippe Fraisse have suggested that the wine bought for the Eileithyaia may have been local. They may well be right: it is identified only as in two inscriptions (ID 440A71, 62, 66; 445.16, 4, 9–10) that specify Knidian and Koan wines (as or and or

[95] IG XI 2.154A15; Glotz, Journal des Savants 11 (1913): 20, for restorations in IG XI 2.144A30, 199A22; Larsen, 392. On Delian viticulture in general, see Bruneau and Fraisse, BCH 105 (1981): 127–53, 108 (1984): 713–30, Philippe Bruneau, BCH 111 (1987): 339–41. See also the brief treatment of prices for Khian, Delian, Kyprian, and some other wines in Luigi Moretti, Rivista di filologia e di istruzione classica 108 (1980): 41.

[96] Cf. Bruneau, 217–18.

[97] Larsen, 393.

---

) in the same year in the immediately preceding accounts for the Posideia. If Bruneau and Fraisse are right, local wine was obviously no cheaper than Knidian or Koan, and it is hard to believe that producers at Kos and Knidos could survive making wine that sold at 4–5 or 6–8 dr/met but Delians could not.^[98]

More to the point, in my opinion, is the price differential between Knidian and Koan wines. If the

of the Eileithyaia was local, then imported Koan wine was cheaper than either Delian or Knidian even in Posideon, three months after the close both of the sailing season and of the vintage. In the incomplete state of evidence for amphorae on Delos,^[99] only provisional inferences are possible, but Jean-Yves Empereur has recently shown that stamped Knidian amphora handles do not begin to appear on Delos in appreciable quantity until the 140s B.C.^[100] Rarity could then account for the prices of the Knidian purchases. But Delian wine cannot have been "rare," or at least was no "rarer" in the 170s B.C. than it had ever been.^[101] It seems more likely to me that the prices for reflect on the one hand the restricted availability of local Delian vintage and on the other the greater productivity of islands like Kos and Khios. Modern travelers in the Kyklades know how difficult it is to get local vintages, although this is owing in part to loss of vineyards to phylloxera in the 1960s and the abandonment of productive orchards as inhabitants emigrated; in antiquity production must have been much higher. Yet Delos, for all that, was small, and the local product cannot have come near to satisfying demand. Under such conditions, the appeal of a far cheaper, quality wine like the Koan would be self-evident; furthermore, the increase in numbers of stamped Rhodian amphora handles on Delos from 220 B.C. on strongly suggests the increased availability of yet another desirable but cheaper foreign vintage. The difficulty for renters of the vineyard estates, I submit, was not a "price decline" but "cheap foreign competition."^[102]

[98] Bruneau and Fraisse, BCH 105 (1981): 141. The Eileithyaia took place in Posideon (IG XI 2.287A84, cf. Bruneau, 217); wine bought for it and the Posideia must therefore have been subject to similar, if not identical, market conditions.

[99] See chapter 5, pp. 163–64, above, and chapter 7, pp. 263–64, 266, below.

[100] J.-Y. Empereur, BCH 106 (1982): 224–25. For Tenos, see Etienne, 217–18.

[101] Vines on the estate of Nikou Khoros were extirpated between 180 and 178 B.C. , but this was a unique case: the destruction occurred after the renter Akhaios son of Zelomenes defaulted by failing to post new guarantors. There was no change at all in the rent it fetched, but since rerentals after such defaults often fetched the same rent, it is far from certain that the presence or absence of the vines had any impact (ID 445.16–24).

[102] For Rhodian handles, see again Empereur, BCH 106 (1982): 224–25.

---

This competition could also have been quite local. As with other foodstuffs, Delian production certainly failed to cover local demand. Given an annual consumption of about 150 liters per adult, the Delians would have needed very roughly 262,500 to 787,500 1/yr. If 10 percent of Delian land were under vines, the annual yield would have amounted to about 84,000 1, or 11–32 percent of demand. Sacred estates on Rheneia might have added another 34,000 1, for a total of 15–45 percent. Kykladic neighbors probably made up the bulk of the shortfall: a recent calculation suggests that a single Tenian vineyard in 1950 produced 460,000 1, which could have covered much or all of ancient Delian demand. Ateliers that manufactured amphorae have been identified on Paros, Naxos, and Antiparos, and Columella knew Kykladic vintages.^[103]

There is further evidence to corroborate this view. From 290 to 250 B.C. (see fig. 6.1), Delos was relatively isolated: a small market, dependent mostly on its close Kykladic neighbors and its own production. The winegrowing Group II estates enjoyed local demand for their product, which probably made them popular with renters; some competition among renters would then account for the gradual, steady rise in rents. Trouble arose, as in so many other areas, around 220 B.C. Mean annual firewood and pig prices rose about 40 percent over former levels, a permanent sitonia fund was established, and many other indicators of a changing economic scene are evident.^[104] Probably in these years—although we have no evidence until the 170s—foreign wine began to make inroads on Delos. It may be that religious conservatism is to blame for our lack of evidence: tastes for and availability of Koan and other wines may have developed from the 220s on, but the hieropoioi may have resisted the new vintages for the Posideia until the 190s or so. It is perhaps not just coincidence that the first attestation of Koan wine for the Posideia comes soon after the Rhodians established their new Island League, through which they dominated the Kyklades. Unlike the earlier league, this one counted Delos as a member.^[105]

There is one more consideration. Group I estate rents enjoyed a recovery after 220 B.C. ; I have suggested, on the basis of rising pig prices, that

[103] Etienne, 219 n. 56; A Tchernia, Le Vin d'Italie (Paris, 1986), 26, 200, 360, for consumption and production estimates, cf. also Amouretti in Agriculture in Ancient Greece, 83 n. 17; Jean-Yves Empereur and Maurice Picon, BCH 110 (1986): 495–511, 647–53; Columella, Rust. 1 praef. 20; Plin., Nat. hist. 14.6.54.

[104] See chapter 4, pp. 111–14, and chapter 5, pp. 171–81, above.

[105] Etienne, 101–24; see further chapter 7, pp. 264–67, below. On inferring local taste from amphora remains, see Yvon Garlan in Trade in the Ancient Economy (Berkeley, 1983), 30–32, and DHA 8 (1982): 145–52.

---

renters may have begun to turn to herding to take advantage of these price rises. That in turn would have rendered Group I estates more desirable. It would have been far easier to expand herding on these estates than on any Group II estate, since Group I estates were essentially vacant land,

, suitable either for barley culture or for stock. Despite the presence on most Group II estates of sheep shelters,^[106] which indicates some herding, renters could not run stock through vineyards, which they were obligated to maintain. If the late third and early second centuries were a period of higher stock prices, renters of Group I estates would have enjoyed an advantage. This may also have contributed to the decline in rents for Group II estates.

The rent histories of the estates of Delian Apollo have proven a rich vein of economic information. Rents of estates were tied to the movement of prices of barley and olive oil, a very satisfying result. Since these prices were set in a local market, rents too must have been determined in light of local conditions. This result makes good sense, given the restricted range of trade for the goods in question in all but the most exceptional circumstances. We have also seen that the varying character of the estates and products they yielded helped to determine the way they responded to changes in the economic scene. Sometimes, however, changes resulted not from economic considerations, but from local social and political preferences. The expulsion of the Athenians in 314 released an enormous pentup demand for estates, which propelled rents far beyond anything justified by economic conditions; to correct the situation, the Delians passed the hiera syngraphe, with its stringent procedures against defaulters. Finally, we have yet another piece of evidence for important economic restructuring on Delos in the last third of the third century. We shall return to these matters in chapter 7.

The Rents of Sacred Houses

09Along with the estates, Apollo also rented out a number of houses. They included buildings used as a smithy, by a porphyry processor, and for other productive purposes; a shop; andrones; and probably at least a few residences. Some, like the houses that used to belong to Episthenes, came into Apollo's hands as a result of the confiscations of 375 B.C. ; others were dedicated to the god by private persons, like the house of Stesileus, which was apparently beyond repair and sold. Others may have passed to Apollo by

[106] Kent, 299–301.

---

testament, and some were probably confiscated from defaulting renters of estates in the 280s.^[107]

By 207 B.C. , the houses were rented on five-year contracts. For the earlier parts of the century, the evidence is less secure. IG XI 2.226, which has new leases (see A11), probably dates to 257 B.C. That date would correspond to the five-year cycle attested later. The date of IG XI 2.268, which also has new leases, has been disputed; Molinier thought he could set it in 267 B.C. exactly, but because of doubts about the five-year cycle during this period, Dieter Hennig argues only that it fell sometime in the 260s—in 268 at the earliest.^[108] One objection to holding that the five-year cycle was in force earlier than 257 is the absence of new leases in IG XI 2.158, which dates to 282 B.C. and so should have fallen in a new lease year. But this is not decisive. The inscription is not complete, and there is no reason why new leases could not have been recorded on the lost portion of the stone (cf. ID 366A94–99) or, as sometimes happened, on a separate stele (see ID 399 and 400, both of 197 B.C. ).^[109]

The real problem with postulating a five-year lease on the same pattern before 257 B.C. comes with a new lease period that must have fallen between the rents paid in IG XI 2.158A15–23 (282 B.C. ) and 161A16–24 (279 B.C. ). The house formerly of Antigonos, rented by Antigonos himself in 282 B.C. for 30 dr but by Arkhepolis in 279 for 60 dr, is still in Antigonos's hands in 281 B.C. For Hennig this is the decisive evidence that a five-year lease cycle based on years ending in 7 and 2 is impossible for the early third century.^[110]

The form of the entry at IG XI 2.162A39–40, however, suggests a par-

[109] Molinier, Maisons, 48. Hennig, Chiron 13 (1983): 447 n. 77, 444 n. 72.

[110] IG XI 2.158A20, 161A20–21, 162A39–40, cf. Hennig, Chiron 13 (1983): 446–47. Molinier, Maisons, 48–50, postulates a four-year cycle based on 285–281 and 281–277; objections at Hennig, 447 n. 78.

---

tial default:

("under the arkhon Kharmos [280 B.C. ], 30 dr, which Antigonos son of Timokrates paid, he owed in addition as part of the rent under the arkhon Glaukiades [281 B.C. ]").^[111] Apparently Antigonos had owed an additional 30 dr () of the rent () for 281 B.C. , which he paid in 280 B.C. But this would mean that the rent in 281 B.C. was more than 30 dr, therefore proving a new rental in 282 B.C.^[112] His rent in 281 may well have been the 60 dr attested for 279 B.C. The presence of a new renter by 279 B.C. would indicate that Antigonos had failed to renew his guarantors or perhaps had died and his heirs had either failed to provide guarantors or decided not to take up the lease.^[113] Thus there is no reason to doubt a five-year rent cycle in the early third century.

There is no direct evidence attesting to the manner in which the houses were leased. The automatic renewal with a 10 percent increase (the

) available to renters of estates did not apply to houses.^[114] Sylvain Molinier presumed that sealed bids were accepted on the model of the estates. Hennig suggests that the rents could also have been preset by the hieropoioi, but his argument, based on occasionally drastic rises in rents, has no force, since the same phenomenon appears among the estates.^[115] No hiera syngraphe for the houses survives; whether one existed is debatable, although IG XI 2.226A11 certainly proves the use of some kind of contract. The only question is whether there were also general regulations, as for the estates.^[116]

Although the accounts preserve records of literally dozens of houses, the information is much harder to use than that for the estates. The hie-

[112] Compare the case of Apemantos son of Leophon, who rented the first house of Episthenes for 51 dr, of which he apparently paid only half in 279 B.C. (IG XI 2.161A22, D69–72), but retained possession (162A17–18, 278 B.C. ).

[113] Both common problems; examples at Hennig, Chiron 13 (1983): 447–51; on the responsibilities of heirs, see Hennig, 449, esp. n. 84; Molinier, Maisons, 52–54, argues that heirs had to take up the lease. Hennig, 451, doubts the annual renewal of guarantors, in my view on insufficient grounds. I thank my colleague A. D. Macro for discussion on these matters.

[114] Hennig, Chiron 13 (1983): 443, 442–43 n. 71; Molinier, Maisons, 50–51.

[115] Molinier, Maisons, 42; Hennig, Chiron 13 (1983): 443–44.

[116] Hennig, Chiron 13 (1983): 441–43, dubious; Molinier, Maisons, 41–42, assumes general regulations.

---

ropoioi were not always scrupulous about distinguishing buildings; sometimes they recorded rents one year for several andrones that were obviously different buildings, but without indications that could help us identify them with the same buildings in other years. Houses also sometimes disappear from the record, only to recur years later. The "houses where Ephesos has his shop," for instance (IG XI 2.158A21:

), occurs regularly in accounts from the 280s to 268 but is mysteriously absent from the fully preserved account of 250 B.C. , only to crop up again in 219 and continue down to the end of our records. It is possible that such gaps represent periods when a house was not rented out because it was undergoing repair, or for some other reason; but these lacunae do not help establish a consistent set of data.^[117]

It is possible to estimate, in a general way, what proportion of the rental property on Delos was in Apollo's hands. In 250 B.C. , the hieropoioi recorded income from the city of 1,690 dr, attributed to a 10 percent tax on rents, the

. Since a 5 percent surcharge was added to the tax (the ), the actual total tax collected was 1,605 dr 3 ob, representing a total rent paid of just over 16,000 dr. Assuming that the monies turned over to the hieropoioi represent the total tax, and that the tax was levied on all rents, the rents Apollo collected that year for his houses—934 dr—account for just under 6 percent of all rents.^[118] Since not all buildings belonging to Apollo appear in this account, the actual total percentage must have been higher, but probably no more than about 10 percent of the whole rental stock (as represented by money rents). Apollo may have been the chief landlord on Delos.

From all the houses recorded from time to time in the inscriptions, I have selected eleven to analyze. Rents for these buildings occur from the 280s to 179 or later, giving a full range of data, with the inevitable gaps. Full details appear in Appendix IV. As in other cases, I have constructed an indexed rent for each house on a base year, but unlike the commodities or the estates, the houses did not provide enough data points in 250 B.C. for a base year. Instead, I selected rents recorded in 269 or 268 B.C. —both from the same rental period, 271–267 B.C. —as the best available compromise. Since it was not possible to distinguish among uses of the houses as neatly

[117] E.g., Ephesos's house, Orthokles' house, and Pythas's, all absent from IG XI 2.287A34–39 but attested in years prior and subsequent to 250 B.C. (see Appendix IV).

[118] IG XI 2.287A9–10, 34–39.

---

---

as among the estates—those in the sample appear mostly to have been used as commercial enterprises—I have aggregated the data to produce a single average indexed rent for each possible rental year (table 6.15).

Inspection of figure 6.8 suggests that rents in general climbed over time. The figures for 247 B.C. and 197 B.C. , however, look clearly out of line with the rest of the data. A model that makes house rents dependent on time, with dummy variables to pick out 247 and 197 B.C. and a correction for auto-correlation that appears in the initial regression (a poor Durbin-Watson of 1.15), yields very good results (table 6.16), accounting for fully 80 percent of the variation in rents.

Several questions arise. First, it is not immediately obvious why 247 and 197 B.C. should be so out of line with the other years. Second, the trend line of figure 6.8 suggests a closer periodization of the data. From 287 to 277, rents clearly rose; the same is true from 212 to 192, with the gross exception of 197 B.C. Between 277 and 212, however, except in 247, rents remained remarkably steady. They varied by no more than 12 percent from the average (104.76) in 262 B.C. , and in every other year by less than 6 percent. What could account for these phenomena?

Let us consider first the exceptions, for which several explanations are conceivable. For both exceptional years we are dependent on only five rents for the aggregate. Perhaps the houses whose rents survived were somehow unusual? This however does not seem likely. Three other aggregates (287,

Figure 6.8
Aggregated Indexed Rents of Apollo's Houses, 287–182 B.C.

---

262, 212) depend on only four to six rents, yet show nothing unusual. For three houses (1, 2, 9.1), 247 gives the highest rents ever recorded; for the others (5, 6) the only higher rents occurred 40–55 years later. In 197 B.C. , two houses fetched their lowest rent ever (4.1, 5), and two others their lowest rents in 80 and 90 years (1, 4.2). Further, rents for most individual houses in 247 and 197 run against the trend for the years before and after. The exceptions are not the artifacts of scarce data.

Commodity prices showed some upward pressure in 247 B.C. Firewood prices were extraordinarily high, pig prices elevated compared to 250 B.C. (but that was an exceptional year), and oil prices too rose. For at least the latter two goods, these adjustments were returns to more typical price levels after the cheap year of 250. But perhaps additionally something was happening on Delos to raise firewood prices radically and to affect rents too. What it was, however, I am at a loss to say.

For 197 B.C. there may be an explanation. That year the Greek world, including the Aegean, was embroiled in the Second Makedonian War. Philip had taken the Kyklades in 201–200 B.C. , but lost them in the summer of 200 to the Rhodians. Naval operations in the Aegean included an attack on Andros, held by a Makedonian garrison, which resulted in the

---

temporary expulsion of the population and the delivery of the island into Pergamene hands.^[119] More immediately important, exactly in 197 B.C. , the Akhaian general Theoxenos, commanding 1,000 infantry and 100 cavalry, passed through Delos on his way to or from assisting the Rhodians in Asia Minor; he stayed long enough to dedicate a silver phiale (ID 425.11, 442B67–68; Livy 33.18.5). If Theoxenos had commandeered housing for his troops, as often happened in the Hellenistic world,^[120] this might well have discouraged potential renters. At the end of the war, Rhodian interests in the Kyklades led to the reestablishment of the Nesiotic League, whose headquarters were set up on Tenos (not on Delos, as formerly), and to the imposition of military responsibilities on the Kyklades. In the aftermath of war and the creation of a new Kykladic hegemon, renters may have hesitated to commit themselves to rents at levels previously accepted. But five years of experience with the new situation convinced renters that economic circumstances had not changed, and rents resumed their climb.

What can we say about the long-term trend of rents? First, house rents have provided yet another indicator of an improved situation after ca. 220 B.C. Like firewood and pig prices and the rents for Group I estates, house rents rose—and strikingly—in the last quarter of the century (more specifically between 212 and 207 B.C. ). The rise in rent occurred rather later than that for prices and land rents for Group I estates, which indicates that whatever was fueling the expansion worked first in those areas rather than on house rents. The inference is important, for it speaks against seeing international trade as the engine of the expanding economy. If prices for pigs and firewood were rising because of expanded trade, the impact ought to have appeared early in house rents, since merchants would have needed buildings on Delos both as warehouses and as residences. Inasmuch as the merchants would generally have been metics, they would have been compelled to rent, and the rising demand they created would have appeared early, not late, in house rents.

The rise in rents may be connected to the rise in firewood prices. I argued in chapter 5 that the wood price regulation the Delians passed prob-

[119] See, generally, Will II , 149–60, with references. On Andros, see Livy 31.45.2–13, with John Briscoe, A Commentary on Livy Books XXXI–XXXIII (Oxford, 1973), 153–54; Theophil Sauciuc, Andros (Vienna, 1914), 83–87. The notice in one manuscript of Livy 33.30.11, taken from Valerius Antias, that Paros was given to the Athenians in 197 B.C. has been universally rejected: see IG XII 5.2, testimonium 1348, Otto Rubensohn, RE, 18.2 (1949), s.v. "Paros" 1824. Briscoe, Commentary, 307–08, does not mention the variant. Etienne, 114 n. 48.

[120] See Launey, 695–713. Cf. also IG XI 4.1030.2–4, Labraunda III 2.46.7–9.

---

ably in the 230s or 220s may have provoked the rise in firewood prices seen from the last third of the century. We do not know for a fact whether the costs of wood for construction rose too, but it seems likely. Higher prices for building timber, which was used for roof beams in Delian houses, would certainly have led to higher buildings costs; that in turn would have put upward pressure on rents.

A different explanation may be offered for the rise in rents from 287 to 277. I argued above that the imposition of the hiera syngraphe for the estates in 300 B.C. led to a series of defaults and confiscations of property of renters; Delians who could not find guarantors under these circumstances were locked out of estate rentals. The exploitation of estates was an economic activity intended to turn a profit. Delians who could not rent them were likely to have looked elsewhere for economic opportunities; they may have found them in the sacred houses, which provided space to carry out all sorts of activities, as we have seen. And indeed, five men who rented one or more sacred estates in the years 314–290 also rented houses in the years 290–240 B.C.^[121] The houses were substantially cheaper than the estates, ranging in the 280s and 270s from 25–55 dr (except house 8, 60–76 dr), and most were under 45 dr. The leases ran only five years, exposing renters to only half as many opportunities to default. The commercial businesses the houses housed were certainly less liable to uncontrollable fluctuations (like the weather) than agricultural enterprises. Best of all, the hieropoioi never seem to have proceeded against defaulters or cancelled their leases, but rather merely to have written them up with the rest of the debtors who defaulted on loans.^[122] Houses provided an altogether safer investment. The flight of some former estate renters to them in the 280s could then account for that decade's rather substantial rise in house rents until an equilibrium was achieved.

With one known exception, that equilibrium lasted until the second-to-last decade of the century. Nothing serves better to indicate Delos's eco-

[121] Ampheas son of Aristeas, Soloe-Korakia in 314, IG XI 2.287A36; Aristeas son of Aristeas, Hippodromos in 303, 199A8 (cf. Hennig, Chiron 13 [1983]: 475, no. 1); Aristeides, Rhamnoi in 297, 158A17 (Kent, 323, no. 40; Molinier, Maisons, 94, no. 24); Autosthenes, Dionysios in 301, Nikou Khoros in 303, 297, 161A19, 162A18 (his son Alexibios also rented a house, cf. Hennig, 479, no. 15); Makhon son of Praximenes, Kharoneia in 314, 203A25, 204.29. See Appendix IV. Other possible candidates include Aristodikos son of Aristokrates, Soloe-Korakia in 307–306, 290.26; Demeas, Pyrgoi in 301, if he is the Demeas son of Eumedes at 203A26, 224A21, 226A20, 287A37; Nikandros son of Xenomedes, Kharetia in 307–303, 287A35.

[122] Molinier, Maisons, 66–68; Hennig, Chiron 13 (1983): 452.

---

nomic stability throughout the period. The arrival and departure of several hegemones, troop movements, and wars, while skewing or distorting prices in the short term, as we have seen in chapters 4–6, failed in the long run to disturb the character of the economy. Especially interesting is the absence of any impact of the end of Ptolemaic hegemony. House rents continued at the same levels as before—as did other indicators. Changes beyond the immediate sphere of the Kyklades had virtually no permanent impact on prices and rents on Apollo's birthplace.

---