You would have noticed in the double-or-even-quadruple-knit stockings patent that the description skipped over certain realities of sock knitting: namely, the heels and toes. In response to Lola’s comment that an explanation of heel shaping was missing, I flippantly answered that one could add an afterthought heel, and that in fact there was a patent on it.
In the meantime, Kate wrote to tell me about some 40s-era Patons publications she had obtained that described the “Beehive Aladdin Heel” and the “Beehive Innovation Sock”; the latter she described as having an instep and three sides of the ankle worked flat, stitches picked up for the heel, and the sole and back of the ankle knitted and seamed to the first place; in short, a replaceable heel.
She asked how one might go about finding out whether a patent had actually issued. And thus begins the lecture. So, instead of writing only about afterthought heels, first I’ll write about the type that requires forethought.
If you’re minded to go searching for such things, the most advanced patent offices have online databases that allow for searching their issued patents and their published applications. This means that within certain limitations, you can not only locate issued patents that could potentially be enforced against infringers, but you can also find out what some inventors or applicants think might be patentable. The databases are also (again, within certain limitations) text-searchable, so you can search by patent owner name, or keyword, and that the very least the patent/application title or abstract will be searched along with other bibliographic data. Most countries now publish pending applications eighteen months after filing; the U.S. started doing this significantly later than other countries did (about ten years after Canada, for example). These applications are searchable the way patents are searchable as well. (This is not retroactive; old patent applications that never issued to patent are not publicly available unless something else happened to cause their publication.)
Patents and applications are searchable not only based on text fields, but also according to classfication. When a professional patent search is done to determine what patents or applications might exist in a given field, it is typically (or should be) conducted by searching the most appropriate art classes for that field. Text searches are very unreliable, and are the most common flaw in, er… self-prescribed or amateur searches: just because you call something a toothbrush doesn’t mean that everyone else calls it a toothbrush. So, if you’ve invented an improvement on a toothbrush and you’re trying to figure out if it’s patentable, searching every single published patent and application for “toothbrush” to see if anyone else ever described your idea is not only insufficient, it’s inefficient as well. For starters, the patent drafter may have decided to call a previous invention a “dental brush” or “bristled dental cleaning instrument”, and never used the word toothbrush at all. Plus, searching for a common term like “toothbrush” will pull up a lot of chaff that is probably wholly irrelevant and a waste of time to review. For example, try searching for “knitting” in the abstract of issued U.S. patents. You will find some highly relevant and interesting patents, but not a lot; you will find a lot of machine knitting patents, some of which are interesting to the hand knitter as well; and you will find a lot of patents that don’t actually have anything to do with two sticks and string (or a knitting machine and string), but make legitimate use of the word “knitting” as well. In short, keyword text searches are only good for turning up previously published patents and applications if you know exactly what you’re looking for, and you know that it already exists, and uses those keywords to describe it. (Patent databases can propagate misspellings of inventors’ names, too.)
In addition to the inadequacy of text searching, there are some other issues to keep in mind if you’re looking for something in a patent database:
First, the older the document, the harder it is to search by text keyword. The U.S. patent database, for example, is only searchable by patent number and classification for patents issued before 1976, and not by text (including bibliographic data such as owner or inventor). However, esp@cenet, the database maintained by the European Patent Office, does have bibliographic data for older U.S. patents. The amount of data can vary; it may only be a title, and not the inventor. Similarly, older Canadian patents only provide a title and owner/inventor information, and so the abstracts are not searchable. But in brief, if you search only one free online database for patents and applications, I’d suggest esp@cenet.
Secondly, as mentioned above, patent applications filed within the last eighteen months have not yet been published, unless the applicant specifically requested early publication. In fact, the U.S. even plays favourites: under certain conditions, a U.S. inventor who does not seek patent protection in an other country that publishes patent applications can even request that the U.S. application not be published until patent grant. Also, these publication rules do not necessarily apply to design patents or industrial designs, just to utility patents. If you’re looking for bleeding edge technological development this way, you’re probably looking at a year-and-a-half time lag.
And, by the way, if you are searching for applications or patents for the purpose of trying to determine whether it is likely that you may be infringing on someone else’s rights now or in the future, do not rely on the status of the patent or application as reported online (whether it is still in force, or lapsed, or expired, or dead). Yes, it’s pretty blatantly obvious when an old patent has expired. But databases can store inaccurate information; you might think that an application has gone abandoned for failure to pay the annual maintenance fees that some countries require, but it could be that the online database isn’t updated as regularly as a patent office’s internal database. (In this country, for example, you really need to call the patent office to find out if a fee due last month was actually paid on time.)
Also remember that even the best searches may miss relevant art, even if you do a comprehensive classification search and a keyword search combined. Sometimes the patent office doesn’t classify an invention in the right categories.
Here endeth the lesson.
To find out whether this knitting patent ever issued, the first question to answer is, “where would they have filed?” This being Patons, the likeliest suspects were the U.S., Canada, and Great Britain. The one database that covers all three countries the best is esp@acenet, but instead, I searched Canada first by searching for Patons and Baldwins as owner. (A patent or application can be assigned to another entity after filing, but the Canadian database does update owner information if assignments are registered with the patent office, and I figured chances are the patent application would have been filed in the name of the owner rather than the inventor. And bingo, I found the patent. Its application was filed in 1942, so now I had a framework for searching the U.S. database, if I chose. But I did not choose to do so, remembering that it would be more efficient to search esp@cenet. This I did, and I found the corresponding U.S. and British patents. You can download these patents yourself, of course (the Canadian patent number is 433936; U.S., 2372468; GB, 559487), but I’ve regurgitated the text and drawings of the Canadian patent here (although it has the drawings from the U.S. patent, because they’re cleaner; I think the text of the Canadian and U.S. patents are the same, but I haven’t double-checked). To save you trouble, I put the British patent here [PDF], because interestingly it includes explicit directions for knitting the sock and replacement heel — a knitting pattern, in a patent. This content did not make it into the Canadian or U.S. versions. In comparing the text of the patents, I think that the patent application was initially prepared by a Canadian or British agent; when redrafted for filing in Canada and the U.S., a single version was prepared in which the unnecessary content was cut (free sock knitting pattern!) and the rest revised in accordance with U.S. requirements: there is no “u” in “color” in either the Canadian or the U.S. patent.
This particular method of making a sock with a replaceable heel was invented by a Toronto woman named Nora Jarvis Allen, and assigned to related Patons and Baldwins companies, depending on the country (for the U.S. patent, Patons and Baldwins Inc. in New Jersey; for Canada, Patons & Baldwins Limited, a British company with a Toronto address; and for the British patent, the same Limited company, but with an address in Yorkshire).
Briefly, the description starts out with the first step of the quasi-afterthought technique that can be used not only for sock heels, but also for glove and mitten thumbs, buttonholes, and pockets, and anything that requires severing knitted fabric along a section of a row: namely, the use of a separate length of yarn (preferably in a contrasting colour) to work a portion of a row, so that this separate length can be picked out after the fact to yield two edges with loops to be picked up and worked. (If you behave in true afterthought fashion, you would not think ahead to use contrasting yarn; you’d simply snip a strand in the middle of the row section that you planned to unravel, and you’d carefully unpick the stitches and place them on the needles.)
In the case of this method of working a sock heel, a typical top-down sock pattern is used (cuff to toe); once the top of the heel flap is reached, it is worked flat towards the toe, with selvedges; the last row of the heel flap, though, is worked with the separate (contrast) length of yarn. Once that last row is worked, the main sock yarn is used to continue the sock; however, the stitches are not picked up along the heel flap selvedges to continue the instep. Instead, stitches are cast on to either edge of the set of instep stitches, which are then worked in the round with the heel stitches towards the toe.
In another variant, three pieces of contrasting yarn are cut. After the heel is shaped, one final row is worked across the heel stitches in contrast yarn. Then, stitches are picked up on either side of the heel flap with the other pieces of contrast yarn, and the work proceeds from there. Thus, in this variant, there are three stretches of contrasting stitches. (Actually, there are four if you follow along with the British patent — this variant is the first set of verbose instructions provided — because the British patent knitting instructions suggest that contrast yarn also be used to graft the toe to make unpicking the seam for mending easier.)
To replace the heel, the contrasting stitches are picked out, the heel ripped back up to the beginning of the heel flap, the live stitches placed on the needle and a new heel flap worked in its place. When finished, the flap must be sewn to the instep and grafted to the foot edge. (Do read the patents for detail, and do follow along by reading the knitting instructions in the British patent if you need more than vague hand-waving.)
This appears to be the likeliest candidate for a patent on the so-called “Innovation Sock” or “Aladdin Heel”; in fact, this is the only patent issued to Patons and Baldwin on a hand knitted sock.
These Beehive publications also granted a peculiar licence:
BEEHIVE INNOVATION SOCK. An application for patent has been filed and although Patons & Baldwins Limited welcome the use of this invention by persons knitting socks by hand with Patons & Baldwins’ yarn, other persons should not use or imitate this invention without first obtaining a licence in writing from Patons & Baldwins Limited. (this example is from Men’s Socks by Beehive, Series No. 62)
BEEHIVE ‘ALADDIN’ HEEL. By making the replacement of heels in socks a simple and practical operation, Patons & Baldwins Limited are glad to offer this contribution toward your personal economy. An application for patent has been filed [etc.] (e.g. Hand Knit Socks by Beehive, Series no, 37)
It is interesting that Patons & Baldwin saw fit to file a patent application, and moreover how they thought to enforce their rights. Their goal was likely to prevent competitors from selling patterns embodying this technique, but note the language of the licence: permission was granted to those knitting by hand, using only their brand of yarn. These days, if such a term survived the enforcement laugh test (could you imagine how difficult it would be to pursue your sock-knitting customers for breach of such a term, assuming that there was a binding contract?) the enforcer would likely face antitrust tying allegations from the allegedly infringing defendant-customer. (This does not necessarily mean that a patent would be categorically invalid or unenforceable.) I do wonder if they did ever make use of their patent to at least prevent a competitor from publishing patterns with similar techniques. (In case you’re wondering, the latest of these three patents would have expired in the 1960s.)
You’ll note that although I invoked the “afterthought” concept in describing the technique, this patent describes nothing of the sort: it requires planning. A true “afterthought” heel would require cutting a stitch at the heel, unravelling, then picking up and working the heel. Like in U.S. Patent No. 315,043, the “Art of Knitting Stockings”: literally, part of a row is cut and raveled in a knitted tube at the location of the heel; the live loops are picked up near either raw edge and half-heel flaps are knit, first by working a short portion even then by decreasing one stitch at each end of every other row; then joining the remaining live loops and the selvedges of the heel flaps. In this particular invention, the method was meant to be implemented using a machine for knitting the flat heel flaps that could also work decreases (“narrowing”); the heel was to be finished by either hand or machine. As well — perhaps given the finer gauge at which the machine-made sock was made (the patent for Young’s knitting machine, referenced in this patent, does not give dimensions or gauges) — the stitches picked up after raveling the partial row were not picked up right at the severed edge, but rather three rows (“courses”) away, in order to ensure a neat product. The remaining rows at the severed edge would be raveled and cut during finishing.
Given the machinery suggested for working the heel flaps, the options of working short rows or an afterthought heel in the round were not given — I am guessing that these would have been advantages more readily achievable by hand, although machines for knitting tubular items were available at the time. But it seems to me that those knitters who turn their minds to knitting socks are a resourceful and inventive bunch — note, I’m not one of them, having never knit a sock.