George Carlson





I have a complete set of Handloader magazine, issue 1 to present.  I often read those back issues to pick up information I may have never learned, or perhaps forgot.  In any case, in one of the back articles written around the 15th year of publication, Al Miller stated in an article about .38 Special loads that all wadcutters (WC) were not created equal. Up to that point in time, I had considered all WCs to be pretty much the same. I mean, how many different “cylinders” could there be?  Why would one be more accurate than another?  Obviously questioning his statements (I should know by now not to do that of Mr. Miller) I proposed to a couple of internet chat acquaintances that we test some of our WC designs in similar revolvers and see what happened.  It turns out that three of us had S&W Model 14s (K-38s), although of different vintages.  We figured that these three revolvers, used by three different shooters, should provide some indication of the accuracy of Mr. Miller’s statements.


We decided to make up ~ 25 bullets from each WC mould we could find for each shooter, and then mail the bullets to each other so all shooters had at least 25 bullets from each mould.  We did this and the shooting commenced in the spring of 2004.  Miller was right, all WC moulds and bullets are not created equal, at least in our limited testing.  More on the results will be provided later.  First I would like to provide some information on each of the bullets tested, and the moulds that cast them.  All bullets were cast of straight wheel weights (WW) or a WW + 1% tin alloy.  They generally ran from a BHN low of ~9 to a BHN high of ~13.5.


Bullets & Moulds


Between four of us (three shooters and one interested party – beagle), we had 9 different bullet designs, and one mould that was undergoing some modifications.  The modified mould was/is an RCBS 35-148-WC that I had purchased off of e-bay used.  I am in the process of trying to restore this mould to usability.  Thus, we had 10 designs, with one of the moulds a slight modification of one of the others.  This mould is identified as RCBS 38-148 Mod in the tables and text that follows.


The casting performance of each of the moulds was not really surprising.  In general, the more you pay for a mould the better they cast.  In the moulds we used the SAECO and H&G were the tops, with the RCBS moulds next, followed by the Lyman and LEE.  It should be noted that each of us visually inspected the bullets to be sure they were “good” before we used them, but no attempt was made to weigh them.  There were no obvious rejects in any of these bullets.




Ø      Lyman mould 35891 (4 cavity) – The bullets cast from this mould averaged 145.87 gr., with a high of 147 gr. (+0.77%), to a low of 145.2 gr. (-0.46%).


Ø      RCBS mould 38-148-WC modified (2 cavity) –The bullets cast from this mould averaged 158.26 gr. with a high of 158.9 gr. (+0.4%), and a low of 157.5 gr. (-0.48%). 


Ø      Lyman mould 35863 (2 Cavity) – The bullets from this mould averaged 146.1 gr., with a high of 146.7 gr. (+0.4 %) and a low of 145.8 gr. (-0.2 %).


Ø      Lyman mould 358495 (2 cavity) - The bullets from this mould averaged 147.2 gr., with a high of 147.7 gr. (+0.4 %), and a low of 146.4 gr. (-0.5 %).


Ø      Lyman mould 358395 (2 cavity) - The bullets from this mould averaged 166.1 gr., with a high of 168.7 gr. (+1.6 %), and a low of 164.6 gr. (-0.9 %).  Note that this is a hollow base bullet style.


Ø      Lyman mould 358425 (2 cavity) - The bullets from this mould averaged 114.7 gr., with a high of 115.2 gr. (+0.4 %), and a low of 114.0 gr. (-0.6 %).


Ø      RCBS mould 38-148-WC (2 cavity) - The bullets from this mould averaged 151.6 gr., with a high of 152.0 gr. (0.3 %), and a low of 151.2 gr. (-0.3 %).


Ø      Saeco mould # 53 (2 cavity) - The bullets from this mould averaged 144.9 gr., with a high of 145.2 gr. (+0.2 %), and a low of 144.5 gr. (-0.3 %).


Ø      H&G mould # 50 (2 cavity) - The bullets from this mould averaged 146.1 gr., with a high of 146.6 gr. (+0.3 %), and a low of 146.4 gr. (-0.2 %).


Ø      LEE mould 38-148 (2 cavity) - The bullets from this mould were not sorted by weight.





Before loading any WC’s for the WC test, I tested bullets sized “.357” and “.358” in my M-14.  My revolver clearly showed bullets sized “.358” shot more accurately than those sized “.357”.  I was slightly puzzled after the shooting as I expected the accuracy to be about the same, which it clearly wasn’t.  The groups with the larger bullets were easily 50% smaller than those with the smaller bullets.  Despite my puzzlement, I proceeded to shoot the tests with the “.358” bullets. The data reported below is the data shot with my “.358” bullets.


After shooting the WC tests, I was still curious about the difference in accuracy between the two differently sized bullets.  As a start, I hauled out my calipers and measured the Lyman 35891 and RCBS 38-148 Modified bullets before and after sizing.  The as cast diameters were about .358 for all of the WC bullets measured.  My bullets sized “.358” were actually .357, and the “.357” bullets were .356!  No wonder my results were so pronounced re: accuracy between the different bullet diameters.  The .357 bullets were about the minimum I’d expect to shoot accurately, and I would expect the .356 bullets to lead and show inaccuracy.  Duhhh!  I guess the message here is don’t trust the numbers stamped on your sizing die.  Measure the bullets after sizing and go from there.  You actually want bullets sized about .001 over the largest cylinder throat.  Hopefully that will be larger than the groove diameter of your barrel.  A second thing of note was the BHN of my bullets.  They were about 13.5.  Using the now accepted pressure factor of 1420 psi / BHN required to produce bullet deformation; my bullets needed something on the order of 19,000 psi to produce bullet upset.  This pressure is significantly above that produced in the loads we tested.  Consequently, any imperfections in the bullets re: diameter, diameter smaller than a chamber throat, or out of roundedness would produce gas cutting and a loss of accuracy.  You can see these phenomena in the chamber mouths, throats, and forcing cones of your revolvers.  If the bullets do not upset, and/or do not perfectly seal the chamber throat, forcing cone and bore, you will see a wash of lead in these locations.  Using a bullet of softer alloy will normally eliminate this leading, and will also result in increased accuracy.  I could see a slight lead was in these locations in my M-14.  I intend to cast some bullets of a softer alloy, probably in the BHN range, and rerun this test using the 35891 mould.  I expect the leading to be eliminated, and the accuracy to improve from ~ 2” to something on the order of 1.5” at 25 yards with the softer bullets.


The bottom line is that you should measure your chamber throats, and bullet diameters to be sure you have a compatible fit.



Bullet Alloys


Note that the BHN range we tested was from a low of 9.3 (one bullet) to a high of 13.5 (two bullets) with most of the bullets in the BHN range of 12 +/-.  The direct comparison of BHN (9.3 vs. 12.5) was with the LEE bullet.  It is worth noting that the accuracy did not change with these two BHN bullets.  It is also worth noting that these alloys were essentially WW or WW with a bit of tin added.  The pressures expected in the .38 Special with WCs in the range of 140 to 150 gr., using 2.7 gr. of Bullseye would be in the range of 9,000 psi to a maximum of 10,000 psi.  The softest bullet we tested would not obdurate until a pressure of around 13,000 psi was developed!  Based upon BHN alone, we shouldn’t have expected to see any difference in the accuracy between the BHN 9.3 and 12.5 bullets!


The bottom line is that you should not trust anything in your cast bullet endeavors.  Measure everything, including chamber throat diameters and bullet hardness.  Research the pressure you should expect with the load you plan on testing to be sure you are not using a bullet that is too hard.  If your bullet is too hard to obdurate with your planned load, the bullet must be at least .001 larger than the largest throat dimension in your revolvers chambers, and the revolver must have a groove diameter that is smaller than the smallest chamber throat.  If this latter situation does not exist, you might get reasonable accuracy, but it won’t be the best your revolver can produce, and you should expect at least some lead wash in the chamber throats, forcing cone, and first inch or so of barrel.


Exceptions to the above do occur, but be thankful when they do, but not disappointed if they don’t.  With cast bullets, you can control all of the dimensions that matter, including bullet hardness.




All of the brass we used was once or twice fired, and all was in good condition.  In general, all cases were from the same lot, or at least of the same manufacturer.  As an aside, a couple of us tried some other brass just for giggles to see what effect brass may have had on our results.  We couldn’t find any difference at all with these loads.  It appears that you can mix or match .38 Special brass….it won’t make much, if any difference with revolvers as near as we can tell.  At least this was true with these light target loads.


All of us used a charge of 2.7 gr. Bullseye.  This load seems to be a standard WC target load that has been used for many decades.  Primers were Federal, CCI, and WSP.  If there was any difference in the accuracy in our shooting as a result of using different primers, we didn’t see it.


As an aside, I had a strange experience seating bullets.  My dies are RCBS and I had experienced no difficulty in using these in years past.  Apparently the once fired R-P cases (lot of 2000 cases) I purchased has a small rim turned inward on the case mouth.  If I seated and crimped bullets in a single step, nearly all of the bullets were damaged (lead shaving) during the seating.  I de-burred all of the cases (for a second time) using an inside neck bevel tool and the situation improved, but I still damaged ~ 30% or so bullets during seating.  As I didn’t (and still don’t) have the answer to the problem as of yet, I finally seated bullets ~ 90% into the case without crimping, and then crimped the bullet in a second step.  No bullets were damaged using this procedure. Why??? I don’t know.  It’s going to take a bit more scratching of my nearly bald head before I sort this out.  I have never had this experience with any other cartridge, pistol or rifle.


All bullets were sized either .357 or .358 and lubed with LBT blue or another known good bullet lubricant.


Shooting Technique


All groups were shot at 25 yards, sandbags, with two 6 shot groups fired with each bullet.  If we were interested in finding the tiniest groups, we would have found the single most accurate chamber and fired the groups singly using the same chamber for each shot.  We weren’t interested in smallest groups per se, but were interested to see if any of the WCs tested were as good as, or better than the others.


Each chamber was fouled, and then two six shot groups were fired with each bullet.  In general, all rounds were fired in similar temperatures for each of the shooters.  For example, all of those fired by Pilgrim were in calm weather with the temperature in the 60’s.  The gun was cleaned with a brush after each bullet was tested, and the process repeated again.  No solvents were used in the cleaning, and the revolver was allowed to cool between groups.  No apparent change was normally seen in the groups between the 1st and 2nd groups.  On the other hand, we did not fire enough groups with any one shooter or gun to make any statistically significant observations from a mathematical sense.  On the other hand, some of the WCs appeared to shoot more accurately than others.  In fact, all three of us found the same bullets performed better than the others for each of us.


A couple of observations may be useful here:


Ø      The K-38 shot by Pilgrim required that the chambers and barrel be fouled before groups were anywhere near consistent.  The difference between groups fired from a clean gun and a fouled gun was about 2X.  Make sure you try your revolver both clean and fouled before you make any conclusions.

Ø      Different bullet diameters can make a huge impact on accuracy.  Pilgrim’s revolver was tested with both .357 and .358 bullets…NOT!  The RCBS bullet sizers were stamped .357 and .358, but were in fact .356 and .357.  The revolver shot reasonably well with .357 bullets, but the .356 bullet accuracy was lousy.  Make sure you measure the bullets you are using and don’t trust what the label says.

Ø      Flawed bullets or bullets damaged in seating will enlarge your groups, but not by as much as expected.  Pilgrim tested this question using Lyman mould 35891.   This test had three parts; perfect bullets, bullets damaged in seating, and bullets with slight flaws in the first driving band.  All of the bullets had perfect bases, including those damaged during seating.  As you can see, perfect bullets gave groups essentially the same as the flawed bullets.  One can only wonder at this point what the results might have been had the bullets been of a hardness so that they might have obturated.


Perfect bullets             1st group 2 ½”

                                    2nd group 2 5/8”

Damaged in seating    1st group 3”

                                    2nd group 2 1/8”

Flawed bullets             1st group 3”

                                    2nd group 2”






Wad Cutter Accuracy Comparison – S&W Model 14-4, 6”





Bullet Weight (avg.)







H&G # 50

146.6 gr.




Saeco # 52

144.9 gr.




RCBS 38-148 Modified 4

158.3 gr.




RCBS 38-148

151.6 gr.




Lyman 35863 1

146.7 gr.




Lyman 35891

145.9 gr.




Lyman 358395 (HB) 2

168.7 gr.




Lyman 358425

114.7 gr.




Lyman 358495

147.2 gr.




LEE 357-148 / BHN 12.5 3

149.6 gr.




LEE 357-148 / BHN 9.3 3

152.6 gr.








1)      35863 needs to be tested further.  The 1st group size shot by Pilgrim using this bullet may have been larger than normal due to a clean chambers and bore.  This has been seen before with this particular revolver, with one or more of the bullets tested.  Strangely, it did not occur with all bullets tested.  Bullet fit perhaps?

2)      358395 may be the best bullet.  Pilgrim had a flier in his 1st group.  The flyer was not expected, and may not be typical.  This design is slow to cast relative to other WC designs; it is a hard to find mould; and the length of time to cast a “pile” of bullets is significantly longer than for the other designs.  This latter fact alone probably makes this bullet of academic interest only unless you only need a few bullets (perhaps for small game shooting or similar pursuits).

3)      The LEE bullets are at least as good as the rest, and better than most.  The moulds aren’t as nice, but…

4)      RCBS 38-148 Modified mould needs further work.  As cast diameters are too large.  This will be reduced by carefully sanding block faces w/ 320 grit paper.  Only a few strokes will be needed on each face to reduce bullet size by ~ .002 or .003” to the needed .358 diameter.  The top of the blocks will also be milled down to ensure perfectly square bullet bases.

5)      The bullet diameters after sizing may not be what you expect, or be correct.  Pilgrims RCBS .358 sizing die sized to .357, and his RCBS .357 die sized to .356.  The net result was lousy accuracy with the .356 bullets, and a nagging question wondering what the performance of .358 bullets might have been in his revolver.

6)      BHN apparently doesn’t affect accuracy.  At least in the range we used.  As noted above, we shouldn’t have been surprised, as the BHN of 9.3 was still higher than the obturation hardness, of ~6 or 7 needed for the pressures produced in our loads.

7)      Perfect WC bullets can be very difficult to cast (visually perfect bullets) unless the mould is of exceptional quality.  The SAECO and H&G moulds produced essentially perfect bullets with every cast.  The RCBS moulds were next re: rejects followed by the Lyman moulds.  I did not record data on the LEE moulds, but with their performance, one should expect a fairly high percentage of good bullets from these moulds.


Back to the Casting Fellows Homepage