For as innocuous as it may appear, an engine's air cleaner performs a number of functions you don't hear too much about. And because of the influence it can have on flywheel torque, let's spend a few minutes discussing a couple of power-related functions that have nothing to do with "cleaning" inlet air.
In particular, we all know that when stripped to their elements of operation, carburetors are basically "pressure differential" devices. That is to say their fundamental ability to meter fuel is largely a function of pressure differences between atmospheric and inlet manifold. Ideally, fuel-bowl vents would always like to experience atmospheric pressure, but that's not always the case. And when they don't, the remainder of a given carburetor's "calibration" can become upset. The same applies to main fuel metering systems. Fluctuating inlet air pressure can be problematic to several fundamental functions of a carburetor.
We also know that the traditional path of inlet air is from a plane that's roughly horizontal to one approaching vertical as it enters a carburetor's throat. This transition can be both difficult and a potential opportunity for "shaping" inlet air. Ignore both these possibilities, and power can be affected . . . typically negatively. And while we're considering the importance of how air enters a carburetor, keep in mind that dynamic pressure distribution around the circumference of the throat is seldom (if ever) uniform. For example, depending upon how you've either chosen to build an air inlet system, chosen a commercially-available system, or been required to do so by some rule, chances are, the highest entry pressure can be found coming from the engine compartment's firewall . . . as derived from the base of the windshield.
Plus, keep in mind that the shape of the cleaner's top (lid) can also come into play, at least to a degree. In fact, see if this visualization helps you understand what's being suggested here. Consider taking a vertical slice (section) through a particular air cleaner assembly. Now imagine what you see in a profile section of the top and bottom with a portion of the cleaner's filter element in place. Now remove the element, leaving simply a profile of the top and bottom of the assembly. What you now view is what the incoming air sees in terms of how it will be influenced when turning from the aforementioned "horizontal" to "vertical" as it enters the carburetor. OK?
Now, let's take the next step. Imagine what you're looking at is a cross-section of an inlet port with the lid profile representing the port's roof and cleaner base profile the port's floor. Put on your "porting" hat and consider what you'd do to improve "turning" the air as it passes through this simulated "port" and into the carburetor throat. If you're about to turn the page back to Bolles' always-excellent column on suspension bits, that's fine, but let me share something before you do. I've personally modified some very popular and widely used air cleaners, using the concept I'm trying to walk you through here, and found upwards of 6-8 horsepower in the process . . . sometimes depending on the initial air cleaner's original design. In fact, I know of one company who marketed one of these designs for circle track applications and more than one experienced engine builder found power over the best previous version they'd been using.
Now I'll be more specific with an experiment you may want to try. On an airflow bench, install your favorite carburetor, air cleaner stud, and top and bottom of the air cleaner assem-bly . . . minus the filter element. Space the lid the same distance from the base that it would be positioned if the element was in place. If you're using a low-restriction element (a reasonable assumption), it will not influence this exercise, anyway. So leave it out. In fact, it'll have marginal impact on how the cleaner assembly performs in application, regarding what you'll be measuring with this experiment.