"Wanderwords" are more than just a category marked miscellaneous. They are the great travelers of Eurasia. The crop up in many languages, seemingly from out of nowhere. They often resist our best efforts to date their age. Some examples, like *kar(r)- have remarkable phonological homogeneity across language families. Some words, like cannabis (not yet included), are travellers of astonishing breadth; that word in particular stretches from Mongolian to Turkish to Caucasian to Indo-European to Semitic to Basque families (!), covering an enormous geography. And yet in no family is cannabis phonologically at home, the true mark of a great Wanderwort. 

These examples tend to defy placement into language families. What is the difference, say, between Iberian substratum *(a)ur- "water" (see Iberian Peninsula Languages) and European Wanderwort SIL?. For one, *(a)ur- is phonologically comfortable in Basque and Iberian - Basque in fact is infamous for the "wandering-a" (see Basque entry under *pl(o)undh- for another example). Thus, we have two very strong candidate languages that may have been the source. For two, *(a)ur- is geographically very confined to Iberia while SIL? stretches a thousand miles of terrain. Any definition of a Wanderwort is going to be loose, there's no way to keep a tight definition without unfairly excluding strong examples.

*kar(r)- "stone," Western European Wanderwort. A generally accepted paleo-IE root. Curchin 2008; Trask 2008; Kroonen 2014.

Cara(e), a town.

Caraca, a town. Modern Carpentani.

Careni, a demonym.

Carietes, a demonym.

Caristi, a demonym.

Caronium, a town. Modern Galicia.

(h)arri, "stone." Basque.

*kark-ú- "pile of stones forming a sacrificial mound." Pre-Proto-Germanic. See entry in Pre-Proto-Germanic. 

*kar-w- "stone." See entry *kar-w- in Iberian Peninsula Languages.

[*pl(o)undh- "lead" Western Europe. Part of the mystery is that lead historically was not from Western Europe; it likely came from the Caucasus through Anatolia and the Balkans. Greek molybos (with */m/ competing with */b/ in later Greek forms, though */m/ is attested in Mycenaean) was borrowed from a Proto-Anatolian *marwidos from which Lydian mariwdo came.

     *plundh- "lead" Pre-Latin. Boutkan 2013.

          plumbum "lead" Latin.

     *pl(o)udh- "lead" Proto-Celtic.

     *bel(a)un Proto-Basque. Trask 2008.

          ber(a)un Basque.

     *bu:ldu:n Proto-Berber. The Berberic forms are mysterious but unlikely to be native to the Berber languages. They are most likely loans from Europe but the exact trail is unknown. Metathesis of *-n seems unlikely but unavoidable unless it's a Berber morpheme. Resolution of this problem will take a Berber expert.]

SIL? "silver" Wanderwort. Most metallurgical names are travelers of the European continent yet with unknown origins. The traditional European name for silver was *h2erģ-nt-o "the white (metal)." No one has been able to make heads or tails of a reconstruction, however. Note that silver was mined first in the Caucasus and then traded westward. Neolithic silver mines in western Europe were not begun till much later, which implies that the name for silver should have traveled from west to east -- yet it doesn't (Fortson 2011) Beekes (2014) attempts to rectify this problem by proposing Greek and Caucasian names for iron are connected with SIL? and underwent some sort of semantic shift. Polome 1989; Trask 2004; Kroonen 2014.

[śalir unknown meaning. Iberian. Occurs in pecuniary formula iltiŕta-śalir-ban. Laminal /s/ is shared with Basque zilar as /śVlVr/ and otherwise unique to Europe. While Iberian and Basque languages are unlikely to be related, they shared a numeration system, the product of years of close contac. Ending -ban evidently mirrors Basque ban "one," "a," "single," yielding something like "one śalir (silver coin?)." Other European words could be re-interpreted as unwittingly internalizing *sila-/*salir- "silver" + *bar "one" as a single word.]

*silabur- Proto-Germanic. 

silubr Gothic.

silfr Old Norse. [Was *-ur- re-analyzed as nom. case -r in Old Norse and later eliminated due to analogy?]

silofr Old English.

seolubr Old English.

siolfor Old English.

siluƀar Old Saxon.

sil(a)bar Old High Germanic

silver Middle Dutch.

silab/Pur Celtiberian. Competed with native Proto-Celtic *arganto-. Other Celtic tongues did not pick up the word.

σῐ́δηρος "iron" Greek. Usually not included by linguists, but Beekes (2014) notes the phonological similarities are too great to ignore. [If there is a connection, root *sil- may have been akin to "the white" or "the bright."]

*sirabras Proto-Baltic. Not immediately connected to alternate Baltic form found in Old Prussian.

sidãbras Lithuanian.

siraplis Old Prussian.

*su/irebro- Proto-Slavic. 

zido "iron" Udian. Most metals were first mined in the Caucasus and the mountains were a chief source of metal (Mallory & Adams 1997). The similarities with Greek σῐ́δηρος has not escaped the notice of Beekes (2014), who links the words to European words for silver. 

*zirar "silver" Proto-Basque. 

zilar "ibid." Basque, standardized.

zirar Old Bizkaian.

zidar Basque, various dialects.

zildar Gizpuzkoan. Cannot be reconstructed further. Evidently related to *zirar somehow. Trask 2004.