Berkeley's reply to these worries concerning scientific truth is to point out that the materialists are in no better position than he to give us scientific truth. Scientific knowledge, according to Berkeley, Locke, Descartes, and all the other philosophers of that time, meant knowledge of necessary connections: in other words, in order to understand how A caused B, for instance, you would have to understand how A could not failed to have caused B. But Locke himself claims that such knowledge is not available to us concerning the natural world, because we cannot understand how the microstructures of objects give rise to their macroscopic properties (in part because we cannot observe these microstructures, and in part because there are no such necessary connections to even be observed in the case of secondary qualities). Since the materialists cannot discover necessary connections, they cannot get to any deep scientific truth. All that any of us, materialist of idealist, can do is to empirically gather information about what observable qualities are constantly conjoined to what other observable qualities.
This reply of Berkeley's is inadequate on several levels. First of all, contrary to what Locke claimed, scientists have been able to discover necessary connections in nature: the identification of heat with molecular motion is one example of such a discovery. Second of all, whether or not it is true that scientific knowledge in the strictest sense depends on necessary connections (which is itself a dubious claim) it is certainly not true that science needs to uncover necessary connections in order to tell us what the world is like. We think, for instance, that Darwin's theory of natural selection tells us what the world is really like, even though this theory involves no necessary connections, but only probabilistic connections. Similarly, though we have yet to discover the necessary connections lying behind genetics, we think that that field has already gone a long way toward telling us what the world is like. Another worry we might raise for Berkeley is this: most of us tend believe that it is possible that science could require us to postulate the existence of in principle unobservable entities: entities that we know exist, because of their explanatory power, but that we will never be able to observe, no matter how advanced our microscopes become. (For a while, physicists believed that neutrinos were such an entity, though later they did find a way to observe them.) On Berkeley' system, we could never posit such an entity, because an object that cannot be perceived is an object that does not exist.
In Berkeley's De Motu, he discusses Newton's laws of gravitation. Like Newton himself, Berkeley did not think that Newton had discovered a new force in the world called "gravity". Instead, he thought that he had discovered certain mathematically expressible laws that we could apply to the objects around us in order to predict their behavior. "Gravity" on this view, is just shorthand for some mathematical equation. Berkeley, we can assume, would treat in principle unobservable entities in the same way: not as really existing things, but as useful fictions which help us to make predictions. This is a view that certain scientists and philosophers of science hold today, but most people find it very unattractive, and so might find Berkeley's theory unattractive for requiring it.