日本蓄音器商會（簡稱日蓄）是最早將唱片工業引進臺灣的日本唱片公 司。從1914 年到1943 年為止，它總共發行了兩千多面的臺灣唱片。雖然其 原盤完整保存於大阪的國立民族學博物館，而其唱片實體和目錄也有不少存 於唱片收藏家手中，但是有關其產製和發行過程的資料卻相當匱乏，以致過 去以來的研究大多是針對其產製發行的結果而非其過程。 筆者有幸於2012 年取得三份新出土的日蓄臺灣唱片檔案資料，包括原盤 清單、錄音日誌、新譜發賣決定書，為我們研究其產製和發行過程提供了珍 貴罕見的一手資料。本文即以此三份新出土的資料為主要根據，輔以其他史 料（金屬盤、聲音檔、報紙、照片、圓標、樂譜等），考證各個錄音的日期和 地點，重建其產製過程，並藉此詮釋日蓄臺灣唱片因應殖民地有限條件的產製策略。 筆者發現日蓄臺灣唱片的錄音地點並非以日本為主，而是歷經了臺北 （1926–1927）、臺北╱東京（1929–1931）、東京╱大阪（1932.3–1933.7）、東京 （1933.8–9）、東京（1934.9–12）、臺北（1936、1938）、東京（1939.5）等七個階 段，且臺灣與日本的錄音有明顯的不同。臺北的錄音多為傳統樂種和樂人， 內容較為多元，演出人員較多，西樂演奏水準較低，錄音音質較差。東京╱ 大阪的錄音，則樂種較為集中，以流行歌和歌仔戲為主，演出人員較少，且 多為固定班底，西樂演奏水準較高，錄音音質較佳。由此推測，錄音地點的 選擇主要是以錄音內容和音質作為考量。由於臺灣欠缺西樂的演奏和編曲人 才以及錄音技術和設備，因此如果演出的內容牽涉較多西樂或要求較高音 質，則須移師日本錄音，但是若牽涉較多漢樂，則在臺灣錄音。錄音時程和 人員的安排則有成本考量。為了節省成本，錄音是採分批進行，每批先密集 錄音，累積存貨，再慢慢發行，且錄音順序不等於發行順序（其中流行歌的 錄音主要集中在1933 年8–9 月、1934 年11–12 月、1939 年5 月三批）。同 時，為了精簡人員，演出人員必須流行傳統兼演，且依樂種和商標的不同而 用不同藝名發行，形成日蓄臺灣唱片的特殊現象。

Nippon Phonograph Company (Nitchiku) was the first among Japanese record company to import record industry into colonial Taiwan. From 1914 to 1943, it issued more than 2000 sides of Taiwanese records. Even though the metal discs of these records have been well preserved at the National Museum of Ethnology in Osaka, while the released records and their catalogues in the hands of record collectors, there remain very few documents about their production and release. Consequently past research on these records mostly focused on the products rather than the process. This paper uses three newly uncovered archival documents of Nitchiku to reconstruct the dates and locales of its recordings of Taiwanese music and to speculate on the strategy behind their production. These three documents include the recording ledgers, the recording log reports kept by the recording engineers, and the release orders. They are supplemented by information gained from the metal discs of the recordings, sound files, newspaper reports, photos, record labels, music scores, etc. Analysis of these diverse sources reveals that the recordings did not take place in Japan only but actually went through several stages, including Taipei (1926–1927), Taipei/Tokyo (1929–1931), Tokyo/Osaka (1932.3–1933.7), Tokyo (1933.8–9), Tokyo (1934.9–12), Taipei (1936, 1938), Tokyo (1939.5). The recordings in Taiwan and Japan exhibit obvious differences. The recordings in Taipei mostly featured traditional genres and musicians and hence were more diverse in their contents and involved a larger number of performers; moreover, the performance level of Western music was lower and the sound quality poorer. In contrast, those made in Tokyo/Osaka mostly featured pop songs and Taiwanese Opera sung by a core group of singers, with a higher level of performance of Western music and better sound quality. These findings suggest that the selection of the locales depended on the contents and desired sound quality of the recordings. When the recordings involved more Western music or required better sound quality, the recordings had to be done in Japan because of Taiwan’s lack of musicians and arrangers of Western music and its poorer recording facilities and techniques. In order to reduce budget, the recordings were usually done in large batches (with the hundreds of pop songs recorded in three batches, namely August-September 1933, September to November 1934, and May 1939) and then released gradually in bits and pieces without following the original sequence of recordings. Moreover, in order to economize the manpower, each of the core group of performers had to be versatile enough to perform both pop and traditional genres, and, in doing so, the same singer was given different artistic names according to the genres and labels. This practice was unique to the recordings of Taiwanese music, as a way to adapt to the limited musical and technical resources of the record industry in colonial Taiwan.