This summary is just from my reading of the article (The rediscovery of first wort hopping, by Freis, Nuremberg, and Mitter, Brauwelt IV:308, 1995; copy supplied to me by Andy Walsh) and is not meant to be comprehensive; but I hope that is useful to some of us. Any errors in understanding the content of the article are mine. I am sure we will all have lots to talk about on this subject; I'm only trying to provide what the original article had to say (in Readers Digest form). I'll do it in sorta outline form.
1. Introductory material. First wort hopping (FWH) was used extensively at the start of the century but mainly in order to enhance bitterness rather than aroma. It was recognized that the higher pH of the *wort* (as opposed to later in the boil) had a positive effect on utilization, combatting the effects of losses from coagulation on break material. The higher pH of the first runnings enhances isomerization of alpha acids. Other attempts were made to actually hop the mash (!!); other early efforts involved running the sparged wort through a hop filter--a "hop front" instead of a hop back, I guess...DeClerk steeped the hops in 50C water before adding to the wort (to remove "unpleasant" stuff); a later worker used 70C water. Both reported enhanced aroma qualities.
2. The experiments. Two different breweries produced the test brews that make up the subject of this article, Pils types. The two breweries make a slightly different version of Pils. At each brewery, the FWH beer was brewed with a reference beer alongside. The FWH and Reference beers at each brewery were done under controlled conditions, identical ingredients, pitching rates, etc., and differed only in the way they were hopped. In both test breweries, hops were dumped into the boiler once its bottom was covered with wort; no stirring--they just sat there while wort was sparged on top of them. Brew A (total hopping: 13.0 g alpha acid per hectolitre of cast wort) was first-hopped with 34% of the total amount added--Tettnang and Saaz that were typically used in aroma additions at the end of the boil under normal conditions. Brew B (total hopping: 12.2 g alpha acid per hl wort) used only Tettnang, but 52% of the total hop amount was used as First Wort Hops. No aroma hopping was done in either brew.
3. Tasting panel results: the FWH beers were overwhelmingly preferred over the reference beers in triangular taste tests (i.e., each taster was given three beers, two of either the reference beer or the FWH beer, and one of the other, and had to correctly identify which two were alike before their preference results were incorporated in the database). 11 of 12 tasters of each beer preferred the FWH beer. The main reasons given for the preference: "a fine, unobtrusive hop aroma; a more harmonic beer; a more uniform bitterness."
4. Analytical results--bitterness: The FWH beers had more IBUs than did the reference beers. Brew A: Ref beer was 37.9 IBU, FWH beer was 39.6 IBU. Brew B: Ref beer was 27.2 IBU, FWH beer was 32.8 IBU. This should come as no surprise, since more hops were in the kettle for the boil in the FWH beers than in the Reference beers. Prior to fermentation, the worts from both breweries showed the following features: the FWH wort had substantially more isomerized alpha acids, but less non-isomerized alphas. This was particularly true of Brew B, which had a higher proportion of first-wort hops. Nevertheless, the bitterness of the FWH beers was described as more pleasing than the (slightly weaker) bitterness of the reference beers.
5. Analytical results--aroma: For the aroma compounds, very distinct differences were measured (gas chromatography) in both the identities and concentrations of the various aromatic compounds between the FWH beers and the reference beers. Because the precise nature of the effects of aromatic compounds on beer flavor are very complicated, it cannot be said with certainty just why the various measurements resulted in the overwhelming tasting preference, but clearly something is going on here. Even though the reference beers had higher *absolute amounts* of most of the aroma compounds, again the FWH beers got higher ratings for overall pleasure.
6. Final comments: each brewery needs to experiment with its own setup for determining what sort of first-wort hopping is best for it. But the alpha-acid quantity should *not* be reduced, even if one gets more bitterness than one would get in the usual way. The tasting panel results seem to indicate that the bitterness in the FWH beers was fine, and mild--i.e. there is little harshness that can appear in a highly bittered beer. If the hops are reduced to compensate for the extra IBUs one gets from the first-wort hops, then the whole benefit of doing it might be lost. The recommendation is to use at least 30% of the total hops as first- wort hops--basically, this means adding the aroma hops as first-wort hops rather than late kettle additions (at least for my setup, and I suspect for many others' too).
That's my quick 'n' dirty summary. I found the article quite readable, aside from the parts where the technical info is too far afield for me to make much sense of it (e.g. the gas chromatography results). Hopefully this will give a baseline that interested readers can refer to for what will undoubtedly be a fairly extensive discussion of this topic.
One quick comment: Bob McCowan mentioned, quite correctly, that the above commentary applies to infused beers--in decocted beers, comparatively little break is formed in early part of the boil, so one needs to consider this. If I read the Brauwelt article properly, infusion beers were the only ones being discussed.