I was not aware that seawater temperatures of slightly below 32°F (0°C) would move the metal (SA-36 plate?) into the brittle zone, but I could be wrong.

I have heard of workmanship stories like this: Instead of doing proper root and subsequent weld passes, welders under time pressure would lay a few rods into the bottom of the weld, and then just do the upper passes. This would leave unfused gaps at the root of the weld, creating a stress riser in addition to the reduced weld thickness.

I don't know the thicknesses of plate involved, nor whether preheat/postheat would be required. (And maybe some of this was not well known at the time, anyway).

Until the early 1980s (and maybe even now), some of the old Liberty (and Knot) ships were still in use as seafood processing plants in Dutch Harbor, Alaska.

It would be interesting to know more of the whole story.

Check out UCS-66 in ASME Section VIII. I can't remember if SA-36 is curve A or B, but both of these have high minimum design temperatures. If it's curve A, anything above 0.5 inches thick is below brittle fracture temperature in that water.

Very interesting. I suspect the main reason for failure in Liberty ship hulls was the time element. Shortcuts were taken and inspection may have been inadequate. Welders with little experience were probably pressed into service. You have heard of the record times from keel landing and launch in just a few days. Not much QA could be done in such a short time. The wrong material may have entered into the problem and may have been the only material available at the time. I'm sure the care taken in the construction of a warship would not have been used on Liberty ships. Liberty ships were "quick and dirty" and would not be subject to close scrutiny. I'm sure the "welding rod filler with weld bead over" was used and certainly would cause failure in a heavy sea.

I worked as an inspector back in the 60's and would fail welds made in 1//2" steel plates on nuclear submarines. Welders would fail to remove slag between multi-pass welds and I would report it and make them grind out all the weld and do it over again. They were not working at a Liberty ship time frame, but still didn't do the job right.

As I understand the history of the problems with Liberty Ships, there were instances of substandard workmanship, but the primary issues were: 1. By increasing steel thickness, the state of stress changed from plane strain and through-thickness toughness (where the material is inherently more brittle) governed the fracture mechanics of the steel plates. 2. The change to welded construction allowed cracks to run greater distances without being interrupted with catastrophic results. With riveted construction, cracks could be arrested (at least temporarily) in shorter distances. Thanks for the interesting discussion. PJ

Sometimes the construction was also a factor, for a few liberty ships sank at their moorings!

You forgot to mention the main cause of Liberty Ships sinking: U-Boot