KEIJL GO (2004) Swifts Apus apus in Amsterdam ... But how many?. LIMOSA 77 (4): 121-130.

In summer 1994, breeding sites of Swifts were surveyed in parts of Amsterdam. This city possibly holds the largest colonies in the country. All parts of the city built prior to 1940 were visited, as well as parts of the city built after 1945 (c. 2600 ha in total, Table 1). We looked for occupied locations by slowly bicycling through all streets in this area between 18.30- 23.00 hrs in the period 15 May - 20 July (48 evenings, 140 hours). We did not try to count pairs; when a Swift disappeared into a roof or other object we continued without trying to establish whether more Swifts were using the same site. The survey resulted in a list of occupied addresses for the surveyed area. Exact location where Swifts entered, roof type, rooftile colour and number of stories of the building were also noted.
      At 497 addresses 1556 locations were found occupied by Swifts. Larger colonies (ten or more entering Swifts) were found at only 49 addresses; at 305 addresses only one or two were seen entering. This survey is not complete, although all large colonies were probably located.
      Swifts obviously prefer mansard roofs - a minority roof type in Amsterdam. Thus, thanks to poorly preserved houses with mansard roofs (with a steep tiled lower part topped by a flat upper part), the oldest part of Amsterdam supports the majority of Swifts. The birds enter these roofs mostly under the wooden top board. Roof-tiles are preferred over shingles, probably because they offer more openings and isolate as well as ventilate better. Mansard roofs however are favoured over other roof types even if covered with shingles.
      Since 1971, results of five Swift surveys carried out in (parts of) Amsterdam were published. Unfortunately, the numbers cannot be compared as the survey methods were never clearly described. Nevertheless, Swifts in Amsterdam are threatened, as the preferred parts of town are being renovated or completely rebuilt. Alternative nesting sites, like nest boxes and purpose-designed roof-tiles, are expensive and, in The Netherlands, only marginally occupied. It would be worth trying to build houses with mansard roofs, offering ample space between the tiles and wooden top board and using the preferred type of tiles. As adult Swifts cannot be distinguished from immatures in the field, and a number of behavioural aspects may confuse observers, a reliable survey method is still lacking and, hence, numerical changes are still unknown. The majority of figures given in regional, country-wide and international breeding bird atlases probably result from counting flocks and dividing these by two. This gives an overestimate, as prospecting subadult Swifts are included. Swifts also visit other colonies and are then double-counted, or double-missed. Birds on the nest at the time of counting are not included. It could be more rewarding to count birds visiting the nests during clouded, perhaps even rainy, but warm weather, when large flocks are absent, as was observed during this survey. Instead of estimating breeding pairs by unreliable methods it is possibly more informative to present total numbers without speculating on the number of breeding pairs, in combination with the number of occupied locations, which gives a minimum estimate, is easily repeated and could be used to identify population trends. Common Swift Apus apus

[free pdf] [dutch summary]