In a peer-reviewed paper titled “Evidence of Design in Bird Feathers and Avian Respiration,” in International Journal of Design & Nature and Ecodynamics, Leeds University professor Andy McIntosh argues that two systems vital to bird flight–feathers and the avian respiratory system–exhibit “irreducible complexity.” The paper describes these systems using the exact sort of definitions that Michael Behe uses to describe irreducible complexity:
[F]unctional systems, in order to operate as working machines, must have all the required parts in place in order to be effective. If one part is missing, then the whole system is useless. The inference of design is the most natural step when presented with evidence such as in this paper, that is evidence concerning avian feathers and respiration.
Regarding the structure of feathers, he argues that they require many features present in order to properly function and allow flight:
[I]t is not sufficient to simply have barbules to appear from the barbs but that opposing barbules must have opposite characteristics – that is, hooks on one side of the barb and ridges on the other so that adjacent barbs become attached by hooked barbules from one barb attaching themselves to ridged barbules from the next barb (Fig. 4). It may well be that as Yu et al.  suggested, a critical protein is indeed present in such living systems (birds) which have feathers in order to form feather branching, but that does not solve the arrangement issue concerning left-handed and right-handed barbules. It is that vital network of barbules which is necessarily a function of the encoded information (software) in the genes. Functional information is vital to such systems.
He further notes that many evolutionary authors “look for evidence that true feathers developed first in small non-flying dinosaurs before the advent of flight, possibly as a means of increasing insulation for the warm-blooded species that were emerging.” However, he finds that when it comes to fossil evidence for the evolution of feathers, “[n]one of the fossil evidence shows any evidence of such transitions.”
Regarding the avian respiratory system, McIntosh contends that a functional transition from a purported reptilian respiratory system to the avian design would lead to non-functional intermediate stages. He quotes John Ruben stating, “The earliest stages in the derivation of the avian abdominal air sac system from a diaphragm-ventilating ancestor would have necessitated selection for a diaphragmatic hernia in taxa transitional between theropods and birds. Such a debilitating condition would have immediately compromised the entire pulmonary ventilatory apparatus and seems unlikely to have been of any selective advantage.” With such unique constraints in mind, McIntosh argues that the “even if one does take the fossil evidence as the record of development, the evidence is in fact much more consistent with an ab initio design position – that the breathing mechanism of birds is in fact the product of intelligent design.”
McIntosh’s paper challenges that science must remain at least open to the possibility of detecting design in nature, as “to deny the possibility of the involvement of external intelligence is effectively an assumption in the religious category.” Since feathers and the avian respiratory system show irreducible complexity, he expressly argues that science must consider the design hypothesis:
As examples of irreducible complexity, they show that natural systems have intricate machinery which does not arise in a ‘bottom up’ approach, whereby some natural selective method of gaining small-scale changes could give the intermediary creature some advantage. This will not work since, first, there is no advantage unless all the parts of the new machine are available together and, second, in the case of the avian lung the intermediary creature would not be able to breathe, and there is little selective advantage if the creature is no longer alive. As stated in the introduction, the possibility of an intelligent cause is both a valid scientific assumption, and borne out by the evidence itself.
McIntosh has published other pro-ID peer-reviewed scientific literature, which we have reviewed here.