An ELT Glossary : Consonant clusters
A consonant cluster is a series of two or more consonants occurring at the beginning or end of a syllable, or between them. For example, the word sprite /spraɪt/ contains the syllable initial consonant cluster /spr/ while the word ranks /ræŋks/ has the syllable final cluster /ŋks/; the words fifth spring contain the cluster /fθspr/.
English permits up to three consecutive consonants in initial position in a syllable (as in sprite above) and up to four in syllable final position - eg glimpsed /glɪmpst/. This may cause problems for learners whose L1 does not permit such long clusters - eg Spanish speakers.
Different languages permit different consonant clusters - eg in Polish the cluster /bzd/ exists whereas it does not in English
Notice that we are talking here about consonant sounds - not letters of the alphabet. For example, the ends of the word long or with are not consonant clusters because, although written with two letters, they are pronounced with one consonant sound only : /lɒŋ/ and /wɪð/.
In connected speech, consonant clusters are frequently simplified even by native speakers to make them easier to pronounce - for example, fifths will often be pronounced /fɪθs/ rather than /fɪfθs/, eliding the /f/ to reduce the number of consonants to be pronounced together. This will also occur across word boundaries- eg in the best thing... the cluster /stθ/ is liable to be simplified by elision of the /t/ : /ðə bes θɪŋ/