How are two species with similar phenotypes identified as different? Have two different individuals that were thought to be different species ever been determined to be from the same species ( or one a subspecies of the other's species ) ?
Species are difficult to define. This is because they form over a gradual continuous spectrum via evolution, one species does not suddenly become two, but lineages diverge (perhaps because there is some kind of barrier - like a mountain range - in the way) eventually reaching a point where we class them as separate species. How we define species is debated but the most generally accepted definition is the Biological Species Concept (BSC). Roughly the BSC defines a species as "a group of individuals that do, or could, successfully breed." Other species concepts do exist, some have their merits, but the BSC is most widely accepted and normally taken with a pinch of salt - it's not always possible to perfectly define a species because evolution and speciation are ongoing processes. Therefore, according to the BSC, two groups can be called different species even if the individuals look identical as long as there is true reproductive isolation.
Reproductive isolation can occur because of physical barriers such as rivers, oceans, mountains, environmental change. This is the cause of allopatric speciation. One species could become split when a subpopulation forms on an island (e.g. if humans take them to an island). If enough time passes without migration they will become reproductively incompatible. Another mechanism is sympatric speciation, when two species form without geographic isolation. This could be by the evolution of new behavioural, morphological or physiological forms which create division within the initial meta-population in to distinct sub-populations which over time become more reproductively isolated. Traits related to reproduction are amongst the fastest evolving and can quickly reinforce speciation (e.g. sperm morphology is extremely varied between species, see Sperm Biology: An evolutionary perspective, chapter 3).
The higher up the tree of biological classification you go the easier it becomes to classify things in to distinct groups, e.g. An oak tree is clearly a plant, and a whale is clearly and animal; further towards species definitions a whale is a mammal and a sparrow is a bird and both are animals; and perhaps less clearly, carrion crows and hooded crows are two different bird species in the Corvus genus. (Formerly these two were considered the same species!)
Obviously problems begin to occur when we consider species that reproduce through means other than sexual reproduction - e.g. bacteria which clonally reproduce. Traditionally this has been done by phenotypic information. We are now in the genomics era which opens new possibilities. This article discusses modern techniques by which utilise genetic similarity & difference between cells to define species and they suggest it to be a good method. However, this article seems to suggest that genomic methods need some refining first before we start classifying species with them.
Currently, it appears more pragmatic and efficient to preserve the current species definition than to replace it, because it is serviceable as a first level of screening and current phylogenetic knowledge is too limited for a universal and sound change in the definition.
As for whether two species have been reclassified in to one I know there are examples but Google is not throwing up anything and there's nothing I can remember off of the top of my head - if I remember correctly the male and female mallard were once considered separate species because they look so different but I can't find anything to back this up. Species formerly described as one have been split in to two as well - gorilla's are one such example.