General description:

3cy michahellis in October have plain grey upper-parts and coverts, although the outer lesser coverts and some of the greater coverts show a barred pattern in many cases. The lower tertials may be dark centred. The complete moult in the under-parts and head has been finished. Birds show much streaking on the crown, nape and ear-coverts. It takes about seven to eight weeks to reduce streaking and by mid-winter (January) many michahellis show clean snowy white heads. 
The bare parts are typically less bright coloured than last summer or as in adults, now showing a brown-red orbital ring, often a dark hue in the iris, much black on the bill and still straw-yellow legs. However, as can be seen in the images, the legs appear clear yellow in some birds and some birds have very limited black on the bill, e.g. only a diffuse streak near the tip on the gonydeal angle. It's hard to exclude 4cy birds in these cases. An example of the average September 3cy michahellis can be found in the first image, photographed on October 01 2001. 
The table below shows the average primary moult score for 3cy michahellis in October: about 9.0 at the start of the month and P10 fully grown at the end of the month.
For details on differences between michahellis and cachinnans, see e.g. the article by R. Klein & D. Gruber in Limicola, April 1997.

Moult stage and pace of remiges and rectrices:

The complete moult started mid-May, when 3cy birds dropped the innermost primary P1. By early July, the average score had increased strongly: only P7-P10 are still old second generation and the inner primaries are new. The score for late-September averaged 9.2 in NW France. 
As in adults, the gap between full-grown primaries and remaining old primaries is about three feathers in general, or slightly more as the table in the July section shows. This large moult gap is typical for michahellis, and not found in e.g. Dutch argenteus.
By October, the most obvious moult is in progress takes place in the primaries. At the start of the month, the average moult score show the ninth primary P9 fully-grown, at the end of the month the average is near P10.

By the time P4 is fully grown, 2cy birds normally start the complete moult in secondaries and rectrices. The ordinary moult strategy in 3cy birds may be slightly later (starting when P5 is fully grown, i.e. primary moult score 30).
Secondary moult starts in early August. The images on the July page showing birds with open wings illustrate that all outer secondaries are still present. By early August, the outermost secondaries (S1 & S2, at the division between secondaries and primaries) are normally dropped individually, but the central secondaries are shed almost as in small groups, creating large gaps of missing secondaries, nicely illustrated by this flying 2cy bird and this stretching 3cy August michahellis. 
The secondary moult finishes with the replacement of the inner secondaries (about S17-S18), close to the tertials (which in fact are secondaries as well). The completion of the secondary moult coexists with moment that P8 is fully-grown. This is late-September or early October in most individuals.

So, the secondary moult is completed in most birds in the first week of October ( finished approximately at moult score 45, about the time P8 is fully grown). An example of an October 3cy bird still showing old secondaries can be found in image 6733. 
The findings in NW France are exactly in line with the findings of Agnar Ingolfsson, who conducted research on Great Black-backed Gulls (L. marinus) and Glaucous Gulls (L. hyperboreus), published in Ibis 112, 1970. The first secondaries to be dropped are the innermost six (what we call tertials), followed (at moult score 25-30) by the rapidly dropping of the outermost secondaries S1 to S15. The two waves meet at S16-S17. In the first week of October, abraded old second generation secondaries can be found in this place: S16-S17, but only in a few birds. The low primary moult score are indicative for (may correlate with) the moult score in the secondaries.

By the first week of October, ordinary 3cy birds have completed the moult in the tail-feathers. The common sequence has been a centrifugal moult: starting with the inner tail-feathers and continuing symmetrically outwards to both outer feathers. This tail-feather moult is completed by September. New feathers are white with various amount of black markings. Some individuals have a clean white tail. The new feathers can be recognised by the fresh fringes and tips.

Complete moult vs. partial moult

In the complete moult during the summer, 3cy michahellis moult all the scapulars, wing-coverts, tertials and flight-feathers. The complete moult started in May, when P1 was dropped and the complete moult will be finished when P10 will been replaced in October. By August, the complete moult in the wing-coverts is finished with replacement of the last central greater coverts and finally the feathers in the carpal edge (outer lesser coverts). As those last lesser coverts are replaced, 3cy michahellis start the partial autumn moult. In this partial moult some wing-coverts and tertials are included. It starts again with the upper tertials and outer median coverts and is nicely illustrated by this individual from October 01 moulting P8. The outer median coverts are clearly missing.
3cy michahellis show a much larger individual variation in moult stage than 2cy birds. (For instance: the same individual from October 01 with P7 fully-grown, while the average score for late-September is 9.2). So its hard to generalize on partial moult stage in 3cy in September, and it may be strongly related to the primary moult stage of an individual.
However, partial moult may start by the last week of August in advanced birds. By late-September, the innermost greater, median and lower lesser coverts and the upper tertials have been replaced in the partial moult. The new coverts are adult-like plain grey. The older coverts, moulted in the complete moult last summer, may still show a diffuse barred pattern. Hence, the new median coverts and upper tertials normally look much more adult-like and fresher than the outer greater coverts and coverts in the carpal edge.

The partial moult in 2cy michahellis is often more obvious and explained in the 2cy michahellis section. See Topography Section for explanation of feather tracts.

Tables:

The surveys were done at resting and preening 3cy sub-adults (with at least no white mirror at P10 and at least some obvious second summer feathers).