Short Definitions: The Terminology in the Nestle-Aland Apparatus

al
From Latin alii, meaning "others." Used to note that the listed reading has support from a significant number of other manuscripts but not enough manuscripts to represent even a portion of the Byzantine tradition. It represents more manuscripts than pc but fewer than pm -- perhaps between 5% and 25% of the total tradition. It is not uncommon to find al used to note a reading where the Textus Receptus departs from the Majority Text.
f1
Symbol used in the Nestle editions (and others) for the Lake Group (l). For details of the group, see the entry on the minuscule 1eap.
f13
Symbol used in the Nestle editions (and others) for the Ferrar Group (f). For details of the group, see the entry on the minuscule 13.
h.t.
Symbol used in Nestle to describe the error known as Homoioteleuton, "same ending" (which see).
M
The symbol used in the current Nestle-Aland editions (26th edition and up) for the "Majority Text." (The same Gothic M is also used in the Hodges & Farstad text for the Majority Text, but not in the same way.) It is thus equivalent in concept to the symbol Byz in the UBS editions, or with w in editions such as Souter's. It corresponds roughly with Von Soden's K. It is not equivalent to the Textus Receptus (s).
In the Nestle-Aland text, however, M has an additional use beyond the equivalent in the other texts. It also serves as a group symbol to include any uncited "constant witnesses of the second order." These "constant witnesses of the second order" are witnesses cited for every variant in the apparatus, but whose readings are only cited explicitly when they differ from M.
The "constant witnesses of the second order" are as follows:
Note that some of these witnesses have lacunae; one should be sure to check that they are extant for a particular passage before citing them on the basis of Nestle. Also, some of the "constant witnesses" are fragmentary; this means that it is not always possible to cite their readings explicitly. This is particularly true of 33 (this is one of the reasons why it was promoted to a first-order witness in NA27), but it is also true of 1506, 2344, and 2377, which remain second-order witnesses.
One brief example must serve to explain this.
In 2 Thes. 1:2 (the first variant in the apparatus of that book), the text has patros o[hmwn]. In the apparatus we read
1,2 o B D P 0111vid 33 1739 1881 pc m bopt | txt Aleph A F G I 0278 M lat sy sa bopt (Y pc: h.t.)
That is, the witnesses B, D, P, 0111vid, 33, 1739, 1881, and some versions omit the word; the remaining witnesses include it. Among these remaining witnesses are, of course, the ones explicitly cited (Aleph A F G I 0278), but also the witnesses comprehended within M -- in this case, K, L, 81, 104, 365, 630, 1175, 1241, 1505, 2464, l249, and l846 (1506 is defective here, and we have seen that P goes with the other reading).
Of course, the Byzantine tradition sometimes divides. In this case, the Nestle apparatus cites all witnesses explicitly, and marks the divided portions of the Byzantine text pm.
pc
From Latin pauci, meaning "a few." Used to note that the listed reading has support from a handful of other manuscripts (seemingly not more than about 5% of the total tradition).
pm
From Latin permulti, meaning "very many." Used to indicate a large number of manuscripts at points where the Byzantine tradition is significantly divided. A reading marked pm is the a Byzantine reading without being the Byzantine reading. A reading marked pm probably has the support of roughly 30% to 70% of the total tradition.
rell
Also sometimes rel. From Latin reliqui, meaning "[the] rest." Used in Nestle-Aland to indicate that all uncited witnesses support the reading. In other editions, it may simply mean that the vast majority support the reading. Some may even use specialized notations after rell (e.g. rel pl, "most of the rest").
vid
From Latin videtur or ut videtur. Closest English equivalent is probably "apparently." The siglum of a manuscript is marked vid if the original reading cannot be determined with absolute precision. This happens frequently with the papyri, where individual letters are often illegible. It may also happen in the event of a correction; the original text (or sometimes the correction!) may be partially obscured. It is generally agreed that vid should only be used in a critical apparatus if the manuscript being studied can only support one of several possible variant readings. (In a collation, of course, uncertain letters should be marked with a dot below the letter or some other symbol; letters which cannot be read at all should be replaced by a dot.)
v.l.
From Latin varia lectio, meaning "a variant (or different) reading." Used in Nestle-Aland refer specifically to readings found in the margin of a manuscript and offered as an alternative to the reading in the text. Such readings are most common in Harklean Syriac (where, however, they are indicated by syhmg), but are also found in 1739 (where the reading of the text is indicated 1739txt) and occasionally in other manuscripts (see, for instance, the notes to 1 John 5:7-8, where we find the passage about the "three heavenly witnesses" shown as a variant reading in 88 221 429 636). It should be noted that variant readings are not necessarily better or worse than those of the text; 1739 has some very interesting marginal readings (e.g. Rom. 1:7, 1 John 4:3), but the readings of the text are generally superior; in the Harklean Syriac both text and marginal readings have value; in the case of 1 John 5:6-8, the marginal readings are obviously spurious.