Ideally, for the digitisation of an entire herbarium, the labels for all specimens should be captured, folder by folder and processed respectively. At the same time, all new incoming specimens should also be digitised before entering the collection. However, in the case of financial issues and staff rotations, choosing the specimens prioritised for digitisation could be a more viable strategy, allowing us to digitise small portions of the collection case by case. This requires extra physical handling of the specimens, but it avoids producing an extra amount of files that can remain unprocessed for years or become outdated (e.g. if the specimen has been re-identified).

Volunteers or technical staff with moderate experience in the taxonomy and geography of the region can be involved. The labels and IDs can be photographed using any digital camera or smartphone. In the LWS Herbarium, labels are usually located in the bottom right corner and accession numbers (six-digit IDs, with sometimes omitted zeros at the beginning) are in the bottom left corner of the herbarium sheet. Therefore, only the bottom part of the herbarium sheet must be photographed (Fig. 1). All pictures must have the same horizontal orientation and size - this will save time during their processing as there will be no need to scale and/or rotate them. However, it sometimes happens that labels are located in the upper part of the herbarium sheet. In such a case, the picture of the entire sheet is taken. The same herbarium can store specimens of the same taxon under different names, as accepted names and taxonomic vision have changed over time. Therefore, it is essential to prepare and follow the checklist of predefined taxa and check both the accepted name and all their possible synonyms (Fig. 2).

Figure 1.  

The common view and organisation of the pictures captured during the labels imaging. Examples of horizontal (A, B) and vertical (C) images containing elements of the herbarium sheet required for the databasing: 1 - accession number; 2 - label(s); 3 - field number. Example of the files organisation (D) representing the key elements of the folder with labels images: 4 - readme file with a list of species/infraspecies represented in the current folder and supporting notes; 5 - images of empty table captured after the last specimen of each species/infraspecies helping to navigate between them quickly; 6 - the label images of the first species/infraspecies; 7 - the label images of the second species/infraspecies; 8 - the label images of the third species/infraspecies.

Figure 2.  

The flow chart of imaging the labels of specimens of predefined taxa from a particular region. The herbarium folders are precisely checked for the specimens corresponding to the predefined checklist. All potential synonyms must be checked and respective notes must be made in a checklist and/or readme file. The labels and IDs of all specimens meeting the checklist criteria and doubtful ones (requiring additional examination and/or clarification of the collection region) are photographed. All the specimens that do not meet the predefined checklist are omitted. There is no need to make extra-detailed (excepting doubtful specimens) photos at this stage, but the quality of the resulting picture must be enough to recognise the text. The same illumination and orientation of the pictures are preferable, as it later helps to process the images faster.

The image files should preferably be sorted into folders named correspondingly after the processed taxa (species or infraspecies). However, in the case of intensive digitisation with many different taxa processed daily, placing the files in folders named by working date could be more convenient, with the prospect sorting the files by taxa later or without such sorting at all.

During the bulk photographing of the specimens' labels in the LWS Herbarium, the contrasting pictures (e.g. of an empty table or blank sheet) are made after the last specimen of each species or infraspecies (Fig. 1D). Such contrasting pictures help to quickly navigate through between the image portions and preliminarily estimate the number of pictures for each taxon. Each folder with the multiple pictures obtained in one day must be additionally supported by a text readme file containing the list of processed taxa and any comments on discovered issues.

The LWS Herbarium contains ca. 147,000 specimens, many of which were collected in the second half of the 1800s and in the 1900s by a few dozen principal collectors. These specimens mostly have handwritten labels, some of which are partly damaged. They were primarily written in Polish, Ukrainian and Russian. However, many labels here were still written in Latin, Slovakian, Romanian, French, German and Hungarian. Few recent specimens were annotated in English. Some specimens combine labels in several languages. Therefore, processing of such specimens requires at least basic skills in multiple languages and palaeographics.

The data from the labels at the LWS Herbarium are extracted only by qualified persons directly from the herbarium labels using the protocol depicted in Fig. 3. The person working with primary data must be familiar with the taxonomy of the selected plant groups and the regional geography. Further data cleaning and cross-validation should preferably be realised by another person experienced in general biodiversity data processing. Otherwise, engaging volunteers requires the development of an advanced crowd-sourcing platform with a rating system, as was proposed by Santos et al. (2020). In such a case, different permission and trust levels must be acquired depending on the transcribers' performance.

Figure 3.  

The flow chart of data mobilisation from the herbarium labels and data enhancement. Initially, the data are stored in a Microsoft Excel table. This table is formatted following the DarwinCore standard (TDWG 2025) with the idea to conform to the GBIF Occurence class of the datasets (GBIF 2025b). Preliminary data cleaning is performed in the native environment of Microsoft Excel after each portion of work (e.g. after finishing processing all taxa from one family or at the end of the year). However, such data cleaning is not enough for further data processing. Therefore, advanced data cleaning is realised using OpenRefine software (OpenRefine Community 2025). It is worth noting that, before uploading to GBIF, the data must be validated in consistency with the GBIF backbone taxonomy (GBIF Secretariat 2023). Three other principal plant checklists, i.e. World Flora Online (WFO 2025a), Plants of the World (POWO 2025) and World Plants (WP - Hassler 2025) also have to be taken into consideration. To validate the taxonomy in consistency with WFO, the Name Matching Tool (WFO 2025b) can be applied. Mistakes in georeferencing can be checked in different ways, for example, using QGIS (QGIS Association 2025) or similar software after completing the dataset.

The LWS Herbarium does not use OCR or HTR technologies to parse the data from the herbarium labels due to lack of expertise and the significant variation of handwriting and languages presented on labels. Unfortunately, the test application of Transkribus (READ-COOP SCE 2024) showed insufficient results and we failed to train it. At the same time, the relatively low number of specimens (i.e. 50-100 specimens per day) can be processed by hand. Therefore, we decided to use the traditional data extraction method until we gain expertise in this field. Besides this, new technologies implementing artificial intelligence in data parsing and enhancement (e.g. DiSScover - DiSSCo (2025a) and Hespi - Turnbull et al. (2024)) are actively developed. Probably, in the near future, data extraction from the herbarium labels will be released automatically with high accuracy.

However, we found it helpful for data processing to create an additional list of the collectors (at the moment, it comprises 374 records), which contains standardised collectors' names in English and shortenings following the IPNI database (IPNI 2025), alternative names (e.g. in original language and other languages), years of life, years of research activity, research interests and geographic coverage. There are also provided links to respective biographic data on IPNI (IPNI 2025), Wikidata (Wikimedia Foundation 2025), HUH Index of Botanists (Harvard University Herbaria & Libraries 2013), Bionomia (Shorthouse 2025), VIAF (OCLC 2025) and other online sources if available. The list is supported by photos of respective handwriting serving as a comparison benchmark. Applying standardised names of collectors allows better data navigation, filtering and sorting. Moreover, there is no need each time to type the correct name or verify its correct spelling, especially if it applies diacritic marks. This saves a considerable amount of time.

Pre-imaging preparation of the specimens is a multi-level process that includes different tasks, most of which can be done by technicians, but specialists must still be involved in quality control. Our experience revealed that, even after qualified transcription of the labels, ca. 5-7% of specimens require additional data modifications and clarifications at the stage of pre-imaging preparation (Fig. 4).

Figure 4.  

The flow chart of pre-imaging preparation of the specimens. Based on the dataset developed during the label transcription, required specimens are selected from the collection and placed in a separate working table for further pre-imaging preparation. The initial preparation of the specimen before imaging includes mounting unbound parts of the specimen and labels to the herbarium sheet, packing the small plant parts into the envelope attached to the sheet, restoring labels or preparing new ones and changing damaged herbarium sheets and/or covers. The second part of the specimen preparation includes placing the stamp of the herbarium, checking and stamping (if absent) the ID (specimen accession number), attaching the barcode label, attaching the nota critica with re-identification (if needed) and stamping the sign 'Digitised'. The barcode can be potentially detached, so the stamping or writing the accession number directly on the herbarium sheet is required. If the specimen is missing or has a duplicate accession number, the new accession number is designated to it and indicated in the dataset. At the LWS Herbarium, the barcodes with six-digit accession numbers as locally unique IDs (LUIDs) are applied. Such barcodes are prepared using Zint Barcode Studio software (Stuart 2024), based on the initial dataset before the imaging. They are laser-printed in high-quality mode on the stickers (38.1 mm × 21.2 mm) and adhered to the specimens before their imaging. The application of additional QR codes with GUIDs will be considered for the LWS Herbarium when the common digitisation strategy for all the collections kept at the State Museum of Natural History of the NAS of Ukraine is agreed upon.

Nieva de la Hidalga et al. (2020) pointed out that each image of the herbarium specimen should contain five principal image elements required for proper image processing and quality control: colour reference chart, scale bar, label, barcode and institution name. Besides this, we have found it helpful to mark already digitised specimens to avoid duplicate processing. Special stickers or a stamp 'Digitised' can be applied for this purpose. Preparation and mounting the stickers is very time-consuming, but it benefits from the possibility of removing them in case of need.

Colour reference charts (colour checkers) are generally required for digitisation, including the herbarium specimens digitisation (van Dormolen 2012, JSTOR 2018, Guiraud et al. 2019, Rieger et al. 2023). They ensure the proper adjustment of the colours and white balance in the images. Amongst the most popular colour reference charts for herbarium digitisation are X-Rite ColorChecker Classic Nano (40 mm × 24 mm) and ISA Golden Thread Object-Level Target x1 (235 mm × 25 mm). There are many other options (e.g. Kodak/Tiffen Q13 consisting of two pieces, 203 mm × 60 mm each), which differ widely by the number of colour patches, their organisation (linear or tabular), size and price. When choosing the colour reference chart, it is important to consider two principal aspects: its size and the ability to automatically calibrate it with specialised software (e.g. ImageZebra (2025)). If the chart is too large, it will overlap the specimen. The linear charts can be placed beside the specimen, along its margin. However, the tabular charts are usually quite wide. Therefore, they must be placed directly over the herbarium sheet to avoid the appearance of extensive empty space around the sheet on the image. If the chart is not allowed for automatic processing, extra time will be required to calibrate the images. Application of outdated or low-quality charts can result in incorrect image calibration. For the same reason, applying charts not widely used requires an additional indication of its producer and type. The best option in such a case is to indicate the reference colour values for all patches of the applied chart in CIELAB (L*a*b*) format (Carter et al. 2019). Otherwise, they can be confused and images calibrated improperly. Our previous studies (Novikov et al. 2024b) showed that the ISA Golden Thread Object-Level Target is the best choice since it is less affected by colour degradation. This target is also preferable due to the extended set of grey patches and the presence of an additional resolution test pattern (so called 'convergent line pair gauge', CLP) for evaluating the images' geometric distortion and spatial frequency response.

Scale bar, crucial for estimation of the sizes of the specimen, is present on many colour reference charts. However, many herbaria use branded scale bars placed beside the colour reference chart. Such scale bars can be printed on special transparent non-reflective film (e.g. biaxially-orientated polyethylene terephthalate (BoPET) film DuPont Mylar), which allows them to be placed over the specimen. At the LWS Herbarium, the X-Rite ColorChecker Classic Mini and ISA Golden Thread Object-Level Target charts, each containing scale bars, are used for digitising specimens of vascular plants. Therefore, we do not apply an additional branded scale bar since it requires additional space and manipulations. However, for digitising specimens of non-vascular plants, we use a smaller colour reference chart, Charttu Nano (without a preprinted scale bar), supplemented by an additional branded scale bar.

Herbarium digitisation can require introducing a new specimen ID system, which can either be used in parallel with the existing one or to replace the latter. However, replacing old IDs with new ones can create errors when specimens have already been published and cited in publications with old IDs. Therefore, the new ID system must have unique IDs that do not overlap with the old ones. This can be realised by providing a unique prefix to the new IDs or by applying GUIDs. The image must contain a machine-readable barcode or QR code for better automatisation. Sometimes, specimen IDs can be mistakenly duplicated or missing in the collection. The preselection of specimens, printing and placing the barcodes (Code 128, ISO 15417) or QR codes (ISO 18004) corresponding to specimens' original IDs, as well as identification of erroneous IDs, are time-consuming. Therefore, a new ID system is often introduced when barcoded or QR-coded IDs are printed and placed consequently or randomly on the specimens. A good practice in such a case is using globally unique identifiers (GUIDs) to ensure effective digital data curation and databasing (James et al. 2018, Nelson et al. 2018, Kirchhoff et al. 2018).

Keeping the same order of the elements on the herbarium sheet is useful. It helps to navigate over the specimen elements on its image. At the LWS Herbarium, the specimen label is typically placed in the right bottom corner, while the specimen accession number (LUID) is written or typed in the left bottom corner (Fig. 5). The barcode corresponding to the specimen accession number is attached just near or slightly over it. The stamp of the herbarium is placed in the right upper corner or, if not possible, in the upper centre of the herbarium sheet. Just before imaging the specimen, the stamp 'Digitised' is placed in the upper left corner of the sheet (such stamp can be also placed after imaging the specimen). Notae criticae, if present and possible, are attached near the original label in the right part of the sheet. The colour reference chart, if possible, is placed on the right side of the herbarium sheet just before imaging.

Figure 5.  

The final image of the entire specimen. The key elements represented on the herbarium sheet during the imaging: 1 - principal label; 2 - handwritten or stamped accession number (LUID); 3 - barcode corresponding to the accession number; 4 - herbarium stamp; 5 - stamp confirming the digitisation of the specimen; 6 - colour reference chart (ISA Golden Thread object-level target); 7 - supplementary label (indicating inclusion of the specimen in a local research programme).

After pre-imaging preparation, the imaging can be done by technicians, volunteers or outsourced to a company. This stage is entirely related to image production, adjustment and file organisation (Fig. 6) and almost eliminates the risk of scientific-related mistakes.

Figure 6.  

The flow chart of imaging and image processing. Before imaging, attention is paid to several important steps. In particular, the presence of all key elements i.e. principal label, ID, barcode, herbarium stamp, digitisation stamp, colour reference chart and supplementary labels must be checked on the herbarium sheet. The illumination must be adjusted to indicate preferences and its uniformity along the specimen must be checked. The camera lens must be kept clean. In case of dirt or dust, it must be carefully removed with the specially provided tools (brush and microfibre napkin). The free space availability on the SD card and predefined camera preferences are also checked. The herbarium specimens must be handled carefully, one by one, avoiding mixing and damaging. During the processing of the images, it is important to keep RAW files intact; only renaming is allowed. The images in distribution format (i.e. JPEG) are the objects to further manipulations (i.e. cropping, rotation, colour adjustment etc.). The sharpening of images is not permitted.

The final imaging of the herbarium specimens can be realised using different technical solutions, i.e. scanners (including flatbed and planetary - see JSTOR (2018), Roma-Marzio et al. (2023), Tasenkevich et al. (2024) for examples) and photo cameras (including DSLR and mirrorless - see Thiers et al. (2016), Sweeney et al. (2018), Takano et al. (2019), Davis et al. (2021) and DiSSCo 2025b for examples). Each solution has its pros and cons. For example, scanners are relatively expensive, slow and have a shorter lifespan. Applying flatbed scanners requires additional constructions to revert it and place the specimen under the scanning surface. Such construction is usually bulky (e.g. HerbScan), requiring extra space to hold and providing instead less space to manipulate the specimen. Moreover, the scanner should not contact directly with the specimen to avoid damage to the specimen and pollution of the scanning surface. Master files received with scanners are generally saved in TIFF format and can reach 700 Mb or more, which requires superior storage capacities. However, scanners with CCD sensors often produce images of better quality (especially regarding the sharpness and colour reproduction) and resolution (Waltham 2013, Mehta et al. 2015, Tejas et al. 2022) compared to the commonly used photocameras with CMOS sensors. Photo stations are usually less expensive to construct, modular and better suited to various working spaces. A distantly mounted photo camera provides much more space to manipulate the specimen. Photo cameras produce images faster and can be set to automatically create the master files (generally in RAW format, but sometimes also in TIFF) and distribution files (lossy images in JPEG format). The obtained files are comparatively smaller in size and require fewer saving capacities. However, the quality of obtained images strongly depends on the camera and lens quality as well as on illumination conditions. Resulting images can have different optical distortions and uneven lightness uniformity.

The quality is crucial for the images of herbarium specimens, as they can be used for investigations at different magnifications. Nieva de la Hidalga et al. (2020) ascertained general requirements for the images of the herbarium sheets (both stored and web-distributed) and pointed out that colour reproduction accuracy (ΔE) should not exceed 5 points. Such colour reproduction accuracy conforms to the two-star quality level of the current FADGI guide (Rieger et al. 2023). Our previous studies (Novikov et al. 2023) showed that many herbarium specimens' images provided in the virtual herbaria do not meet this criterion, reaching only a one-star quality level. Most plants lose their natural colours during herbarisation and continue to change their colour over time during storage. Considering this, the value of accurate colour reproduction is questionable. Nieva de la Hidalga et al. (2020) also suggested 72 PPI image resolution sufficient for web publishing and a resolution of 600 PPI - as suitable for preservation and research. As the authors pointed out, such estimates are fair for the scanned images, but hardly convertible to apply in photography. Based on our experience, we believe that 10 Mp (ca. 72-96 PPI) is the minimum resolution of photo-captured images of the herbarium specimens for web publishing and 20 Mp is the minimum limit for research and long-term storage.

Nieva de la Hidalga et al. (2020) suggested three types of files that should be produced as a result of the herbarium digitisation, i.e. archiving master file (TIFF), hi-res production file (JPEG2000) and lossy distribution image (JPEG). At the LWS Herbarium, we apply only two types: archiving master file (RAW) and distribution file (JPEG). The hi-res production files (80 Mp maximum) can be generated later from the master files if needed. Moreover, we provide distribution files in the highest possible resolution (i.e. 16 or 40 Mp), which is enough for most production and research purposes. Acceptance of such a strategy saves time since both file types are automatically generated by the photo camera.

It is worth noting that image sharpening is not allowed for the archiving master files (van Dormolen 2012, Rieger et al. 2023). It is also recommended to avoid sharpening adjustments in the distribution files. The general processing of the files must include only fixing the image orientation (in most cases to portrait), cropping (FADGI recommends leaving some free space around the specimen; Rieger et al. (2023)) and colour and white balance adjustment (using presets generated for the applied colour reference chart following the producer's recommendations).

The publishing stage includes publishing the data, publishing the images and crosslinking the data and images (Fig. 7). It also involves data transformation to meet the standard applied in the Data Centre Biodiversity of Ukraine (DCBU 2025), as it does not follow DarwinCore. Such data transformation is a sophisticated process mostly manually implemented by the service provider or trained technicians. Therefore, it is not described here. In general, final data verification and publishing can be done by technician staff or volunteers with elementary experience in IT.

Figure 7.  

The flow chart of final data verification, cross-linking with images and publishing on GBIF. Special attention is paid to data uniformity and consistency. Data from different rows must not be mixed. Data verification is realised in the OpenRefine (OpenRefine Community 2025) environment following the protocol illustrated on this flow chart. It is also important to regularly check and keep the UTF-8 coding format of the dataset. Data are published to GBIF using the IPT of the State Museum of the Natural History of the NAS of Ukraine. The multimedia files are deposited to the NIRD Service Platform (Sigma2 2025) and linked to the GBIF dataset using the Simple Multimedia extension file. The structure and data represented in the Simple Multimedia extension file are additionally verified in OpenRefine before uploading to GBIF.

Biodiversity data should be published and permanently archived in appropriate, trusted, general or domain-specific repositories (Egloff et al. 2016) and follow the FAIR and FAIR Digital Objects (FDO) principles (Wilkinson et al. 2016, Guizzardi and Sales 2022, GoFAIR 2025). Amongst the important requirements for published data are their persistence, the ability to clearly identify them on the web (e.g. using DOI) and the ability to track the changes. In such a sense, GBIF (2025a) seems to be one of the best solutions for data publishing. However, it is not a specialised virtual herbarium like JACQ (JACQ consortium 2025), Open Herbarium (Open Herbarium 2025) or Reflora (Institute of Research Rio de Janeiro Botanical Garden 2025). GBIF serves as a global data aggregator with outperformed functionality. At the same time, virtual herbaria are primary data providers with their specific functional peculiarities mainly focused on work with images and extended specimen annotations.

The LWS Herbarium does not have its virtual herbarium platform and Ukraine does not have a joint specialised national platform or portal for publishing data on digitised collections. Only two independent Ukrainian platforms allow the publication of biodiversity data, including those obtained from the digitised collections, i.e. the Ukrainian Biodiversity Information Network (UkrBin 2025) and the Data Centre Biodiversity of Ukraine (DCBU 2025). Both platforms publish different kinds of biodiversity data, including living observations, published reports and data mobilised from natural history collections. However, both platforms are self-running and do not exchange data with GBIF, limiting data integration and visibility. Moreover, they do not yet allow bulk data export, which makes data extraction laborious. As a result, data published through these platforms are weakly integrated into the research compared to GBIF (2025a).

Considering this, the LWS Herbarium data are published online on several platforms simultaneously, i.e. GBIF (2025a), Open Herbarium (2025) and DCBU (2025). GBIF is a principal data-providing platform where the data are deposited as Occurrence class datasets through the Integrated Publishing Toolkit (IPT - Robertson et al. (2014), GBIF (2025c)). The images are hosted on the NIRD Service Platform (Sigma2 2025), kindly provided by the Norwegian GBIF node and synchronised with the dataset using Simple Multimedia extension files (Robertson et al. 2014). The Open Herbarium is a free online platform for publishing virtual herbarium collections developed by Arizona State University (ASU) on the basis of the Symbiota software (Gries et al. 2014, Symbiota Support Hub 2025). The images of the digitised LWS specimens are stored on kindly provided ASU servers and synchronised with the Open Herbarium platform. The DCBU serves as the internal platform to publish and host digitised materials, running independently by the State Museum of Natural History of the NAS of Ukraine. Images of the LWS specimens in the highest possible resolution can be accessed through GBIF or Open Herbarium. At the same time, due to technical limitations, the DCBU hosts images of reduced quality (2.25 Mp maximum). These platforms are updated manually and irregularly, without the application of any collection management system.

In archiving, the application of storage media (e.g. optical discs) produced by different manufacturers is advisable to avoid possible manufacturing defects and potentially low production quality (Rieger et al. 2023). When archiving, at least three backup copies must be created using at least two storage media and one of the copies must be stored outside the originating institution (so called '3-2-1 rule' - Perkel (2019)).

Similarly to the case of data publishing, the stored materials must be well-structured, well-annotated, clearly identifiable and represented in open and raw formats (Hart et al. 2016, Wilkinson et al. 2016). There are no clear terms of how long the digitised materials should be preserved, but at least a few decades are expected. Appropriate archiving directly designates the data lifespan. Therefore, it must be realised by archive specialists and cannot be delegated to volunteers or inexperienced staff. De Smedt et al. (2024) recommend cold storage (i.e. archiving without manipulating the files after that) of herbarium images daily. Similarly to the Herbarium of the Meise Botanic Garden (BR), the images at the LWS Herbarium are archived after renaming, processing and quality control. However, at the LWS Herbarium, the images are archived only at the end of the digitisation bench (once per several months). They are stored together with metadata and dataset files describing these images.

Simultaneous use of several different types of storage media (e.g. magnetic tapes and external hard drives) for the herbarium digitisation is advised by Haston et al. (2012b). Application of storage produced by different manufacturers is also advisable to avoid possible manufacturing defects and potentially low production quality (Rieger et al. 2023). Following these recommendations, in the LWS Herbarium, the so-called '3-2-1 archiving rule' is applied (Perkel 2019). This means that at least three backup copies are stored on at least two different types of storage media and at least one copy is stored outside the LWS Herbarium (Fig. 8).

Figure 8.  

The flow chart of the data and images archiving. At the LWS Herbarium, the data are archived on three types of the storage media: (a) on the internal server using the service of the National Academy of Sciences of Ukraine; (b) on the SD memory cards (128 GB SanDisk Extreme Pro cards, which can store ca. 1,024 image files in RW2 format or 6,400 in JPEG format); (c) on the BluRay MABL discs (25 GB Verbatim SL BD-R, which can store ca. 200 image files in RW2 format or 1,250 in JPEG format). Such a combination of storage media (magnetic, electronic and optical) provides long-term preservation of the digitised data and images. Besides this, datasets are also archived on Zenodo (CERN 2025). Since 2024, the data regarding the digitised specimens have been published as printed herbarium catalogues (e.g. Novikov et al. 2024a, Novikov et al. 2024b), stored also on Zenodo (CERN 2025). These catalogues are supported by DVDs (Verbatim Archival Grade Gold DVD-R) with datasets and images in JPEG format. The catalogues with the attached discs are distributed through the libraries that serve as additional archiving agents (Harvey and Mahard 2020, Frick and Greeff 2021).

However, combination of several different archiving media has drawbacks, namely, controlling the preservation conditions and data persistence can be complicated. Moreover, the State Museum of Natural History of the NAS of Ukraine has no common archiving and data management strategy yet. Developing a common archiving strategy would help resolve questions about file formats, priority long-term storage media, cyclic media renewal terms and the distribution of personnel roles in data archiving.