Two very different classifications of the insect Order Thysanoptera have been promoted in recent years. These reflect two very different approaches to classification—phenetic systems that emphasise differences, and phylogenetic systems that search for evolutionary relationships.
The family level classification employed here recognises two suborders, the Terebrantia with eight families and the Tubulifera with one family (together with a second family known only from fossils). Despite being based on phylogenetic principles (Mound et al., 1980), this classification is by no means a fully satisfactory representation of evolutionary relationships (Mound & Morris, 2007). Among the eight families in the Terebrantia, the Fauriellidae and Adiheterothripidae, seem unlikely to represent single evolutionary lineages. Moreover, the two families Merothripidae and Melanthripidae are not distinguished satisfactorily, although their members differ in both body form and biology. Species of Tubulifera, the family Phlaeothripidae differ greatly in structure from all other Thysanoptera, but molecular evidence provided by Buckman et al (2013) indicates that the Tubulifera and the Terebrantia are sister-groups.
In contrast to this system, Bhatti (1994, 1998, 2006) has espoused a radically different, essentially phenetic, solution to the classificatory problems associated with structural diversity among Thysanoptera. That solution involves recognition of two Orders, 10 superfamilies and 40 families. However, many of these families comprise only one, or a small group of a few, species that have some particularly unusual structural attribute. Although entirely logical as a means of filing away unusual-looking taxa, with each named group being precisely defined by some particular structural attribute, this emphasis on differences is essentially antagonistic to the principles of phylogenetic classification. In practise, such a classification can provide little information about possible evolutionary relationships that might be useful to other biologists.
How to classify?
At first sight, the function of a "classification" is to group together objects that are similar, and to separate these from objects that are dissimilar. In biology, however, a classification has far broader objectives, and its essential function is to express in a succinct form the presumed evolutionary relationships between sets of organisms.
The objective of such a classification is to enable other biologists to consider how particular attributes, whether biological or structural, have evolved in their target organisms. For example, has some particular attribute arisen independently in unrelated organisms, such as a "fish-like" body form, or has it been inherited in different species from a common ancestor. If an invasive thrips species is identified as belonging to a group in which most species live only on plants of the pea family, Fabaceae, then it is sensible for us to consider the invader as having the potential to be a pest of lucerne (alfalfa) and related pea crops. Evolutionary relationships are equally important when searching for suitable biocontrol agents for particular pest species. Thus a phylogenetic classification has broad economic as well as academic implications.
Modern classifications in biology are based on the principle that relationships can be inferred only from the possession of shared derived character states; differences, whether in structure or biology, can tell us little about relationships, and thus have little predictive value. Consider a non-biological example, such as the problems faced by some person who has developed a very large collection of domestic tableware. Should this collection be arranged (classified) according to function (all plates together kept separately from cups), or style (patterned plates and cups kept separate from non-patterned), or material (china separate from plastic), or colour? Each of these is entirely logical, depending on the objective of the collector. However, most collectors will classify such objects by their origins (French, Italian or American), or by individual manufacturers. Such an arrangement will emphasise the progressive development of styles, materials and colours, with different producers sometimes borrowing ideas and sometimes evolving their own. This classification thus encapsulates information beyond the mere appearance of the objects. This analogy is intended to emphasise the point that the most generally useful classification is predictive; it has the potential to incorporate richer information than the mere factual differences between the objects classified.
Bhatti JS (1994) Phylogenetic relationships among Thysanoptera (Insecta) with particular reference to the families of the Order Tubulifera. Zoology (Journal of Pure and Applied Zoology) 4 (1993): 93-130.
Bhatti JS (1998) New structural features in the Order Tubulifera (Insecta). 1. Amalgamation of labro-maxillary complex with cranium and other cephalic structures. Zoology (Journal of Pure and Applied Zoology) 5: 147-176.
Bhatti JS (2006) The classification of Terebrantia (Insecta) into families. Oriental Insects 40: 339-375.
Buckman RS, Mound LA & Whiting MF (2013) Phylogeny of thrips (Insecta: Thysanoptera) based on five molecular loci. Systematic Entomology 38: 123-133.
Mound LA 2011. Order Thysanoptera Haliday, 1836. Pp 201–202 In Zhang, Z.-Q. [Ed] Animal biodiversity: An outline of higher-level classification and survey of taxonomic richness. Zootaxa 3148, 1–237. 
Mound LA, Heming BS & Palmer JM (1980) Phylogenetic relationships between the families of recent Thysanoptera. Zoological Journal of the Linnean Society of London 69: 111-141.
Mound LA & Morris DC (2007) The insect Order Thysanoptera: classification versus systematics. Pp 395-411, in Zhang ZQ & Shear WA [eds], Linnaeus Tercentenary: Progress in Invertebrate Taxonomy. Zootaxa 1668: 1-766.