I believe that "legally" the rules say the POH. That being said I go with the latest recommendations from the engine manufacturer. I haven't seen one POH in the airplanes that I fly that have any updated information in them to reflect the realities of the current fuels going into the airplane or the current knowledge available on what's best for the engine. Yeah, I rent old airplanes. Some POH's were written before 100LL, before digital engine monitors and before about 40 years of accumulated knowledge about those engines.
That.
Sometime in the 80's, after I installed a per-cylinder digital engine monitor and started following the actual in-flight CHTs/EGTs, I discovered that the original Cessna-installed single cylinder CHT probe, which was mounted on Cylinder #5 (of 6), was nowhere near the hottest cylinder, ever, and there was sometimes a wide temperature spread among the cylinders (up to 100 deg F in some extreme cases).
This made me realize that by monitoring the Cessna CHT monitor during climb, I had been actually destroying the engine (which in fact happened, I had to replace a cylinder which got warped or cracked).
So I called Cessna in Wichita and asked to speak with an engineering rep, and got through to an old veteran who sounded like one of the design team. I described to him what I had discovered, and he refused to accept it. He kept saying, that's impossible, the engine case is a big heat sink, there can't be more than a few degrees difference between any two cylinders. Well, we agreed to disagree, and ever since then, over thousands of hours, I have been "flying the temperatures": never allowing any CHT to get too hot, and haven't had any engine issue since.
So, anecdotal evidence, but support for the fact that Cessna was not aware of engine reality (e.g. how fuel/air or the cooling flow is distributed to normally aspirated 6 cylinders in real life) when they wrote their manuals and hooked up that single CHT probe.