I think the reason for the Garratt not being accepted in the US can be broken down into narrow gauge and standard gauge reasons.
For US narrow gauge railways, by the time Garratts had been developed past their 0-4-0+0-4-0 initial design, the market for new narrow gauge steam in the US was nearly zero and those that were buying weren't about to try anything new.
For US standard gauge railroads, the features that made the Garratt so successful in the rest of the world simply didn't apply. The Garratt allowed for a deep firebox and a short, fat boiler barrel that allowed for a high powered boiler to be carried on a relatively small loading gauge. In the US, big articulateds could be built with huge fireboxes and combustion chambers which accomplished the same thing.
Garratts also can track and operate smoothly over lines with sharp curves, typical of Colonial railways, but not an issue on US lines that needed the additional power of an articulated loco. Many US standard gauge 2-10-2's developed more power than the largest standard gauge Garratts ever built and without the mechanical complexity of articulation.
Add to this the concern that a Garratt experience a reduction in the factor of adhesion as coal and water are consumed and you can see why US roads would be reluctant to try the design.
The standard gauge New South Wales class 60 Garratt, which looks big, had a TE of 59,560 pounds (later raised to 63,490) but consider that Western Maryland Scenic #734, a 2-8-0 and GTW 6325, a light 4-8-4, both have 60,000 pounds TE and you can see why US roads didn't feel the need for Garratts, at least as they were built and used in the rest of the world. Maybe a Super Garratt would have worked over here, but by then the railroads were buying F's and Geep's.
Michael Allen