To start with a cliché, parti-gyle brewing is one of the oldest tricks in the book. Originally, it wasn’t actually a trick, but an inherent procedure in old brewing methods. The traditional approach was to conduct separate mashes on a given parcel of grain. The first wort would be completely run off, then the grain re-mashed with hot water and the second wort completely run off, and so on for a third, and even sometimes a fourth mash. It was customary to make strong ale from the first wort (sometimes combined with the second), and to produce a much weaker “small beer” from the remaining worts. It seems that this practice may have changed in the first quarter of the 18th century, when porter came onto the English brewing stage. London brewers came round to the idea of combining all the worts from separate mashing so as to make one beer, known as “Entire,” or “Entire-Butt,” and later becoming porter. Ale brewers (that is those producing pale and amber ales) still pretty much stuck to making two or more beers from the same charge of grain. The practice of sparging the grains after mashing and continuous collection of wort to make one final beer is relatively modern, dating from the early 19th century. Even through that century it was common for brewers to make, say a pale ale, and a weaker “dinner ale” from the same batch of malt. Since the total product from a brew was known as a “gyle” (or “guile” in some older books), this technique came to be called parti-gyling.
   
Today, most beer is produced as an entire gyle, but there are a few commercial brewers who practice the parti-gyle approach for certain beers. I believe Harpoon does so for its 100-barrel series, and Woodforde’s of Norwich in England does so, making a 7% ABV barleywine and a 4.3% ABV golden ale from one mash.
   
Why would a modern commercial brewer want to parti-gyle at all? Surely the technology is there now that he/she can produce whatever beer required in one shot, whether big or small? Well, it boils down (and that is a deliberate pun) to commercial demand. If the brewer wants to make, say, a barleywine or imperial stout, but sees only a limited market for it (and remember brewers have to sell their product to make a living), a normal full batch may be more than is required. And in order to get all the extract out of the malt that could mean that a long boil is required to get the wort down to the target OG. This may be uneconomical, or even impossible. But, if the brewer parti-gyles by taking the first runnings of the wort for the big beer, the brewer can hit his or her target gravity right away, with no need for a long boil, and with only a small, saleable volume of the beer. The rest of the runnings can then go to make another beer such as a regular bitter or stout that fits into the brewer’s normal portfolio.
   
So how can this technique help the homebrewer? Well, when it comes to volume, since you may not want to make as much as 5 gallons (19-L) (or 10 gallons/38 L if you’re brewing on an even bigger scale) of a very strong beer. Or perhaps your mash tun just won’t hold enough grain for a full 5-gallon (19-L) brew length of an Imperial IPA. Or perhaps you are limited for brewing time and just want to brew two different beers on the same day without doing two mashes. Or perhaps you are just a homebrewer who likes to experiment (are there any other kinds of homebrewers?) and want to try out a traditional approach. And last but not least, there is another kind of parti-gyling you might want to try. That is to collect a single wort in the usual manner, split it into two, and brew two beers with different varieties or amounts of hops, or different hopping schedules. Such a comparison is a great way to find out which might be the best variety for a particular beer style and how it might best be used to obtain the results that suit your taste. This approach can also be used to compare the performances of different yeast strains. But if you do such experiments take careful notes of what you have done!
   
Parti-gyling is a technique which can offer a great deal of flexibility to the homebrewer as there are other ways than the above in which you can use it. For example, let’s say you do your first mash and collect the runnings for the bigger beer in the usual way. Then, instead of just doing a second mash to produce a beer weaker, but similar to the first, before mashing with more hot water you add in a small amount of specialty malt. You might add chocolate malt and/or roasted barley to turn a pale beer into a brown ale, or a dry stout. Or you might do something quite different by conducting the second mash, collecting the runnings and converting this wort to something else by adding malt extract towards the end of the boil. In this way you might start by making an Imperial IPA from the first wort and then adjust the second wort to give you a “regular” IPA, rather than keeping it as is to brew an ordinary bitter.

(from:byo.com)