LSJ and the Greek dialects

LSJ's entry for πλῆθος is instructive, in that it repeats the lemma as a dialect form:

πλῆθος, εος, τό, Dor. and Arc. πλῆθος Schwyzer 84.8, al. (Argive, found in Crete, v B.C.), IG5(2).6.20 (Tegea, iv B.C.), etc.; Boeot. πλεῖθος ib.7.3171.46 (Orchom. Boeot.); pseudo-Dor. and pseudo-Aeol. πλᾶθος GDI5176.21 (Crete), IGRom.4.1302.18 (Cyme, i B.C./ i A.D.), Hippod. ap. Stob.4.1.93: (πλήθω, v. πίμπλημι):

That is to say that πλῆθος is common Greek, that it came to be written with <ει> in Boeotian (in the Hellenistic period), and that, between the end of the third century BCE and the the turn of the eras, hyperdoric and hyperaeolic forms were used with <ᾱ> for what was always a long e vowel.

It would be simpler and more accurate to give:
'πλῆθος, εος, τό, Boeot. πλεῖθος (3rd c. BCE), hyperdial. πλᾶθος (>3rd c. BCE) [πίμπλημι] ...'.

This neatly illustrates that LSJ simply reports which dialects have which form rather than organising them in accordance with the innovations or divergencies that they showfrom the standard.

Schwyzer 84 provides a neat illustration of a point of which LSJ shows awareness here: the findspot of an inscription does not indicate its dialect. That is determined by certain diagnostic features of the Greek thereon. Although this was found on Crete, its dialect is Argolic. The text is a treat that was brokered between the Tylisi and the people of Knossos, but Argivis mediantibus (cf. no. 83 per Argivos factum, which is Colvin, no, 38 and Buck, no. 85; Anne Jeffery's notes).  

The image in Anne Jeffery's notes for 170.39b seems in fact to be of Schwyzer 84 (B, the central column, beginning δυσμενέας τιθείμεθα καὶ τ|ο̑ν), not DGE 83. It confirms that the script is certainly not green (Cretan), and is in fact blue (B 5 ψάφο̄ν and B 6 μάχᾱ).

LSJ's entry for πῆμα is a similar case:
'πῆμα, ατος, τό (same form in Dor.), ... '.

In other words πῆμα is simply common Greek.

LSJ gives no dialect label s.v. ἔξαστις.
The genitive cited is ιος (Galen uses an acc.pl. ἐξάστιας): its citations (from medical authors) and an inscription from Samos (iv c. BC) are compatible with labeling this as Ionic.

ἐπημοιβός has no dialect label, nor does ἐπαμοιβός. Likewise ἐξημοιβός. However, s.v. ἀντημοιβός we read 'Ep. for ἀνταμοιβός, corresponding, Call.Del.52.'. s.v. ἀντᾰμοιβός, we find only 'v. ἀντημοιβός'.