Now we come to the consonants.
Castilian Spanish is one of the few European languages to include a voiceless dental fricative /θ/ in its phoneme inventory. As this is a very common sound in English, it should present no problems whatsoever for the English speaker. Unfortunately, orthographically, it is either <c> or <z>. This leads non-Spanish-speaking native English speakers to associate it with a lisped /s/, and many will refuse to use it, on the grounds that “it sounds cissy”. Try telling a madrileño taxi driver that he sounds cissy, and see where it gets you! However, certain names seem to have beaten this: Olazábal for one. When the golfer of that name first became prominent, the mispronunciation used as an anglicization put the stress on the wrong syllable, but included the /θ/ correctly (despite the consistently correct stress in the BBC Pronunciation Unit’s recommendation from the day he hit the news, it was only when he insisted in a press conference that this was right that anyone took notice of it. So much for the influence of the Pronunciation Unit).
The voiced counterpart to /θ/ is not listed as a phoneme in its own right, but occurs as the most frequent allophone of /d/: [d] appears in word initial position following a rhythmical pause, and following /n/ or /l/. Elsewhere it is realized as [ð]. Here is another sound that causes no problem to English speakers, but seeing it spelt as <d> makes them shy away from the easy but non-intuitive interpretation.
<b> and <v> are interchangeable – representing [b] word initially following a rhythmical pause, and following /m/; and [ß] elsewhere. Generally speaking, it is acceptable to interpret them into English as /b/ and /v/ respectively, so that Barcelona is anglicized with initial /b/, and Valencia with initial /v/.
There’s not a lot that can be done with /g/ (either [g] or [ɣ]) except treat it always as English /g/. N.B. Spanish <ng> is always [ŋg] regardless of position, and so /ŋg/ in English.
Spanish <ll> varies greatly from one dialect to another, but an English speaker can never go wrong treating it as the sequence /l/ + /j/, so Portillo becomes /pɔ:’ti:ljəʊ/. Likewise the palatal nasal, <ñ> easily becomes /n/ + /j/. Beware of words that look as if they ought to be spelt <ñ> but aren’t: Habanera, the name of the famous aria from Carmen, for example, means “of Havana”, and has no tilde above the<n>, and so should be pronounced /æbə’neərə/ in English.
<ch> represents (almost) the same sound as its counterpart in English: macho, machismo are pronounced ‘mætʃəʊ, mə’tʃi:zməʊ, and not, as so often heard *[‘mæxəʊ, mə’kɪzməʊ], as if they were somehow German or Italian.
<h> on its own is always silent.
<j> and <g> before <e> or <i> are the voiceless velar fricative /x/, so Sergio becomes in English /’seəxɪəʊ/ (or /’seəhɪəʊ/ for those who can’t cope with Scottish or Welsh names).
<qu> represents /k/ before <e> or <i>. The <u> is never pronounced separately in this position.
<x> is usually /ks/, but sometimes, as in México, may be pronounced /x/