Fasteners for wood lath and plaster ceiling

[PSUengineer1]

I have an old building from 1908 with a wood lath and plaster ceiling. I suspect the ceiling fell down (only about 30 square feet of ceiling) due to smooth nails backing out over time. I have an older version of USG Gypsum Construction Handbook, but i don't see specific section regarding fasteners for wood lath to wood floor joists? Can anyone point me to the code reference I am looking for (fastener requirements for wood lath to wood joists)? Thank you in advance.

 

[phamENG]

I would be genuinely shocked if there's a prescriptive requirement for that.

Are you trying to get a feel for the historic assembly, or are you trying to specify a repair to - more or less - match the existing?

 

[PSUengineer1]

I would be genuinely shocked if there's a prescriptive requirement for that.

Are you trying to get a feel for the historic assembly, or are you trying to specify a repair to - more or less - match the existing?

i am trying to figure out why ceiling fell and have at least a modern-day code reference to back up my opinion. my opinion being that smooth shank nails were used to attach wood furring in lieu of ring shank. you know of a code reference i can use for attachment of wood lath to underside of wood floor joists (it can be a modern-day code)? thanks again.

 

[GeorgeTheCivilEngineer]

Early 1900 would be iron nails so they're as likely to have rotted through after 110 years. Can you not inspect the nails and/or the hole they came from?

I assume you mean the modern code requirement but I don't believe there would be one for lath and plaster. Here we'd replace with plasterboard (drywall). If it was heritage then just nails them to each joist as was done originally. I would derive the arrangement from the ceiling that hasn't fallen down.

 

[XR250]

You need to inspect the nails to see if they rusted as George said or they just pulled out. I imagine 100 years of moisture and temperature cycles under a sustained plaster load could cause them to pull out. Why do you need to determine the cause? For me, it would be, "rip all of it out and replace with sheetrock" due to the high liability.

 

[Ron247]

What kind of nails are there, cut nails, smooth wire nails or some other flavor? Cut nails have been the ones I saw more problems with because of the ones that are tapered profile. Pull loose a little is synonymous with pull loose a lot. I would investigate the environmental and physical situation of the area that fell relative to the area that did not fail. Also, are they using the same fasteners? Something made 30 sqft fall but not the rest. I would assume the rest is close to failure unless there are different characteristics of the areas.

Below are 2 responses from ChatGPT about your situation. No idea how accurate the answers are, but they sound ball-park correct.

#1
Ring shank nails—also called annular ring nails—were developed in the early 20th century, with widespread use beginning around the 1930s to 1940s. These nails were designed to improve holding power by adding rings or ridges along the shank, which bite into the wood and resist pull-out far better than smooth-shank nails.


The exact date of invention isn't pinpointed to a single event or person, but their development aligned with the rise of engineered lumber and increased demand for more secure fasteners in construction, particularly in subflooring, sheathing, and roofing.


They became especially common post-World War II during the U.S. housing boom, as building techniques modernized and performance standards increased.


#2

In 1908, the most common method for attaching plaster to ceilings was wood lath and plaster, and the primary design and construction documents of the era would typically reference:

Wood Lath & Plaster Construction​

Lath: Narrow strips of wood (usually ~1/4" thick, 1–1.5" wide) nailed horizontally to ceiling joists.
Plaster: Applied in three coats:

1. Scratch coat – pushed through the gaps between lath to form “keys” that mechanically lock the plaster to the ceiling.
2. Brown coat – for leveling and thickness.
3. Finish coat – smooth surface layer.

Common Design Documentation in 1908​

The most common design document wouldn’t be called a “spec sheet” or “drawing set” as we think of today but would include:

• Architectural blueprints or plan sets (hand-drawn, ink or pencil on linen or vellum)
• Pattern books or builder’s manuals, such as:
  • Radford’s Architectural Plans for Houses (popular around the time)
  • The American Builder’s Companion by Asher Benjamin (earlier but still referenced)
  • Carpentry and Building magazine (published monthly with construction practices and specs)

These documents often noted “wood lath and 3-coat plaster finish” for interior ceilings and walls. More detailed architects might include spacing of lath (often 3/8" gaps), nail types (cut nails or early wire nails), and joist sizing.

Alternatives in 1908 (less common but emerging):​

• Metal lath was just beginning to see limited use in fireproof construction or commercial buildings.
• Rock lath (paper-covered gypsum board) would not become widespread until the 1920s–30s.

If you're looking at a historical restoration or referencing 1908 specs, wood lath and plaster with hand-mixed lime-based plaster would have been the standard, and the most detailed design documentation would come from architectural plan books and specifications written directly by the architect or builder.

 

[phamENG]

To do what you're asking, I'd use the NDS to check the capacity of the nails in withdrawal. Granted, it won't be terribly useful as it will have a margin of safety built in that makes it useless in predicting actual failure. For instance, I could shoot a 2x4 overhead with a nail at each end with a 40lb capacity. I could start doing pull ups and I bet my arms would fail long before the nails come out. (And I weigh considerably more than 80lbs.) But let it hang there for a while, go through several temperature/moisture swings, and the result may be a bit different on another day.

To determine the cause of the failure, listen to George and XR. You'll need a detailed visual inspection of the debris to see what the actual failure mode(s) was/were.

 

[phamENG]

No idea how accurate the answers are, but they sound ball-park correct.

Come on, Ron. ChatGPT and similar AI platforms have their uses, but this isn't one of them. "Sounding ball-park correct" is a pretty low bar for a website for professionals to discuss technical topics.

 

[ANE91]

This is clearly a forensic assignment, but the question misses the forest for the trees. Whether the ceiling was original construction or installed shortly thereafter, you’re trying to frame several decades of successful performance in the context of modern codes. How could you possibly argue a construction/design deficiency? Moreover, you’ll never see ringshank nails in this application; they’re too expensive.

Focus your energy on proximate and contributory causation. There doesn’t need to be a reason beyond “everything falls down eventually.” The overlying framing should be deflected downward under normal load, and that slight curvature exacerbates fastener withdrawal, as plaster is comparatively more brittle. Throw in wood creep, temperature/moisture cycles, and sustained tension, and there’s your recipe for causation. You might even find that the ceiling joists are cracked at unfortunately located knots. You don’t need a document to validate the unwritten laws of physics.

 

[Ron247]

phamEng, the original post asked for a reference, of which I have not seen any offered up. Chat at least gave 3 possible references. I am skeptical of Chat probably more than you are, but that does not keep me from at least looking. When I post any info on this site that comes from AI, I note the source so users can take from it what they may AND I do not edit the information. As far as a low bar, I was only using it to get info like when ring shank nails might have come into play and what was the more common installation methods, not a discussion on how to analyze the issue and devise a repair methodology. For that, I would never rely on AI although I might still inquire. I would have used a more scientific method, but I could not find my OUIJI board or Magic 8 ball. But I fully understand and appreciate your comments.

 

[phamENG]

When I post any info on this site that comes from AI, I note the source

A practice I appreciate - few things more irritating than people posting AI generated blather as if it's their own work. Guess that's why posting any AI stuff gets my hackles up.

 

[jerseyshore]

Whenever I have to check something old I usually start by checking one of the older NYC codes. They are so simple and well written I am envious of the simpler code days (as we discussed in that other thread).

Looks like the 1922 code has a whole chapter on plaster so that would be my suggestion as a place to start if you are trying to figure out what the standard was for the time period.

I agree with pham that it's doubtful you'll find anything useful in modern codes about this specifically.

 

[SWComposites]

I've owned two houses with plaster and lath (1914 and 1925 vintage), and ripped a lot of it out when remodeling. All of the lath was nailed with small smooth shank nails. Some of it came out easily, other areas required lots of prying. A lot depends on the type of wood used for the studs.

 

[Ron247]

Ditto phanEng.

I know that at least 30% of what I ask for is incorrect to date. "AI" at this stage is more of a multiple Google search tool than a creativity tool. Problem is, there is a lot of bad information on the internet that gets repeated by all the people who must write a paper, write book or JUST as bad, respond to an Eng-Tips thread for some reason. This in turn adds another incorrect reference to the internet. AI has no real way of ever sorting these things out. It never throws the bad into a trash can forever, it just repeats it. The following are 3 examples of the abuse it has tried to deal to me already, but I recognized the errors.

The first one I remember running into and STILL see, is that you can "Convert" from mass to weight by some multiplication process. Conversion as I was taught, is feet to inches, inches to mm, or miles per hour to feet per second and you use a Conversion Factor. That is how math teachers told me for years. The converted item never changed its "stripes". Someone wrote on the internet one time that you were "Converting mass to weight", rather than saying they were "Calculating" weight given mass and gravity. I have stumbled into this for 5 years and it is never going to go away. I had an older Chat version use this very statement to imply mass and weight were the same. I have already run into several that use this statement to confirm to themselves that mass and weight are the same thing.

Another is an AI recap of an article. I wanted the definitions of static and dynamic loading. So, I Googled them. AI recapped an article that contained both. The headings and definitions it gave were in the wrong order. I looked up the article, and it presented them correctly. It can't distinguish an article with ONLY 2 options. It also showed a definition that implied dynamic loads only affected pressure vessels. Yeaaaaaaaah, I no longer have to deal with Dynamic Loads on anything other than a pressure vessel.

And lastly, my favorite. I could not remember the formula for the location of the centroid of a ramp load. So, I asked Chat. It gave a formula but stated the distance was from the left. I recognized it was from the right. I next asked for something else related to that request. It gave me my new info but also told me it was wrong on the previous output about left versus right. Had I not asked for the second piece, it would have never told me it was wrong on part of the original question. Gotta Love It.

 

[phamENG]

Yep. In many cases, it's worse than a google search. But I've used it successfully to develop simple python and html applications and to summarize my own technical reports. Of course, I read them afterward to make sure they don't make a mess of it, but it's nice to not have to manually scale down reports to create executive summaries anymore. I've always hated doing that.

When carefully trained and used for well defined purposes, AI models are really powerful and useful. But as soon as we think that it's actually intelligent, we're likely to misuse it. It may well get there, but we're not there yet.

@PSUengineer1 , are you still around?