More reliance is perhaps to be placed on the genealogies of the two royal families given by Herodotus (VII 204, VIII 131). Leonidas (r. 488-480) and Leotychidas (r. 491-469), with whom we are on sure historical ground, are represented as fifteenth in descent from Eurysthenes and Procles respectively. According to all analogies therefore we should expect that the two latter flourished not very long before the middle of the tenth century. In other words the date given by Eratosthenes for the 'Return of the Heracleidai' would seem to be from a century to a century and a half too early. A very reasonable explanation of the difficulty has been suggested by Prof. Meyer who points out that several passages in Herodotus' history seem to imply the reckoning of a generation at forty years[264]. Among the figures given we find (II 145) Heracles dated about 1330. Cleomenes, his descendant in the twentieth generation, was born about 530, or at all events not much later. For Eurysthenes and Procles, in the fifth generation from Heracles, this would give about 1130, which is not very far from the date fixed by Eratosthenes.
It is to be remembered in the first place that the date fixed for Eurysthenes and Procles is apparently that of their birth, and, secondly, that Eratosthenes' scheme is probably only a modification of a previously existing system, other varieties of which are quoted by Prof. Meyer[265]. Indeed it would not require any great exercise of ingenuity to point out traces of a more or less symmetrical distribution of the period covered by the reigns of the early Spartan kings[266]. But, apart from any such speculations, we can hardly doubt, in the light of Prof. Meyer's showing, that the date for Eurysthenes and Procles is derived ultimately from a calculation based on the genealogies rather than from any contemporary written record or tradition. The genealogies themselves of course may represent tradition, so far as they are not interpolated[267], but they point, as we have seen, to a much later date than that which we have been discussing[268]. If we substitute 32 × 15 for 40 × 15, starting from the birth of Cleomenes, we are brought to about the year 1000. That must be regarded as the date really indicated by Spartan tradition for the birth of Eurysthenes and Procles.
We may now turn for a moment to the genealogies of the other Heraclid families, namely those of Argos, Messenia and Corinth. The first of these places Pheidon in the sixth generation, according to one version, in the ninth according to another, from Temenos, the uncle of Eurysthenes and Procles. Unfortunately different dates are assigned for Pheidon. The earliest, which is not generally accepted, places his reign about the middle of the eighth century. But even this, taking the longer form of the genealogy, does not carry us appreciably farther back than the Agiad list. The Corinthian genealogy places the last king, who is said to have been killed in 747, in the thirteenth generation from Heracles. This would agree with the longer form of the Argive genealogy; several of the names however are generally regarded as suspicious. The Messenian genealogy is materially shorter.
Apart from these Dorian genealogies there are some notices relating to the ancestry of persons belonging to other parts of Greece, which must not be ignored. Herodotus (II 143) states that Hecataeus, who was a prominent man at the beginning of the fifth century, claimed to be descended in the sixteenth (i.e. fifteenth) generation from a god. This probably takes us back to the Heroic Age, when divine parentage is common, whereas later it appears to be almost, if not entirely, unknown[269]. Again, it is believed that the genealogy of the Philaidai at Athens, which actually survives, though only in a corrupt form[270], placed Philaios, the son of Aias, in the twelfth generation above Hippocleides, who was archon in 566. Further, according to Pausanias (I 11), Tharypas, king of the Molossoi, who was born soon after the middle of the fifth century, claimed to be descended in the fifteenth or sixteenth generation from Pyrrhos the son of Achilles. It will be seen that, though these genealogies do not agree exactly, the discrepancy is not very great. They seem to indicate the existence of a belief that persons who flourished in the first half of the fifth century were removed by about fifteen generations from the Heroic Age.
On the other hand Pindar (Pyth. IV 9 ff.) in an ode written in 466 and addressed to Arcesilaos IV, king of Cyrene, places that king's seventh ancestor, Battos I, in the sixteenth generation from Euphemos the Argonaut, a contemporary of Heracles. This exceeds even the Agiad reckoning, for Pleistarchos, the representative of that family reigning in 466, was only in the twenty-first generation from Heracles. From the other non-Heraclid genealogies we should have expected that the number of generations to Arcesilaos would be about what is recorded for Battos I[271].
Whatever may be the explanation of this case, it will be seen that the other non-Heraclid genealogies are shorter than that of the Agidai by at least three generations—if we equate Philaios, Pyrrhos (Neoptolemos) and the grandson of Hecataeus' god with Aristomachos the grandfather of Eurysthenes. The dates which they indicate for the 'floruit' of these persons are in no case earlier than the middle of the tenth century. As to the relative value of the two traditions we have nothing to guide us, and the same remark applies to the Greek genealogical evidence in general. Two points however must be insisted upon: (i) that the calculations of scholars of the Alexandrian age, or even earlier times, are not to be interpreted as evidence of tradition; (ii) that the evidence of tradition, whatever be its value, brings the end of the Heroic Age at least towards the close of the eleventh century.
Apart from the evidence discussed above, unsatisfactory as it doubtless is, chronological data for the Heroic Age itself seem to be entirely wanting. We know however that a highly advanced civilisation flourished in the Aegean in early times, and that it was succeeded by a long period in which both art and general culture were at a very low ebb. This latter period, which is commonly known as 'geometrical' from the type of art which prevailed in it, lasted, so far as one can judge, until about the end of the eighth century, at which time oriental influence began to make its appearance. The 'orientalising' period again continued down to the beginning of the classical age. It is a common and natural hypothesis to equate the low-watermark of culture early in the geometrical period with the generations immediately following the Dorian conquest of the Peloponnesos. But unfortunately we cannot thereby obtain any certain date for the latter, since Greece appears to have had little contact with the outside world during the geometrical period.
In recent years some advance has been made through the operations carried out by the British School at Sparta, which is perhaps the most important site for our purpose. From the stratification of the deposits Mr Dawkins, the director, has come to the conclusion that the earliest temple and altar at the sanctuary of Artemis Orthia date from the ninth or even the tenth century[272]. The temple, which must have been one of the earliest known, appears to have been a narrow and unpretentious structure of crude brick and timber. Some geometrical sherds were found beneath the floor, a fact which shows that the sanctuary had been in use somewhat earlier. If the sanctuary was founded at the beginning of the Dorian settlement at Sparta it is obvious that this result agrees well enough with the date indicated for the conquest by tradition. No relics of pre-geometrical times appear to have been found.