PROTECT YOUR DNA WITH QUANTUM TECHNOLOGY
Orgo-Life the new way to the future Advertising by Adpathway-
Loading metrics
Open Access
Peer-reviewed
Discovery Report
Discovery Reports capture exploratory initial findings centered around a valuable and interesting research question.
- Ruike Wei,
- Shangning Yang,
- Xinlan Chen,
- Yicheng Lu,
- Yao Liu,
- Fuliang Hu,
- Stephen J. Martin,
- Luo Zha,
- Huoqing Zheng
x
- Published: November 19, 2026
- https://doi.org/10.1371/journal.pbio.3004091
Abstract
Our reliance on honey bees (Apis mellifera) for pollination services and honey production continues to increase due to the global decline in other pollinators and rising demand for honey. The ectoparasitic mite (Varroa destructor) remains the biggest pest of honey bees forcing beekeepers into continual miticide use. Beekeepers on the Tibetan Plateau, however, report rarely seeing mites in their colonies, unlike all other areas of China. We conducted a series of translocation experiments to investigate this phenomenon. Here, we show that hypoxic conditions above 3,500 m on the Tibetan Plateau reduce mite fecundity and arrest offspring development, and so prevent the long-term persistence of the mite population. This reduced mite fertility and arrested offspring development were also found in honey bee colonies maintained at low altitude but under hypoxic conditions like those found at 3,500 m. This discovery of a physiological limit to V. destructor reproduction highlights a previously unrecognized sensitivity of this mite to reduced oxygen availability, the mechanistic basis of which warrants further investigation. This discovery could unlock new chemical-free innovative mite-control methods that will benefit both the beekeeping industry and help ensure food security.
Citation: Wei R, Yang S, Chen X, Lu Y, Liu Y, Hu F, et al. (2026) Altitude-induced hypoxia suppresses Varroa mite reproduction in Tibetan honey bees. PLoS Biol 24(11): e3004091. https://doi.org/10.1371/journal.pbio.3004091
Academic Editor: Anurag A. Agrawal, Cornell University, UNITED STATES OF AMERICA
Received: May 7, 2026; Accepted: October 19, 2026; Published: November 19, 2026
Copyright: © 2026 Wei et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data is available within the manuscript and Supporting information files.
Funding: This work was supported by grants from the China Agriculture Research System to HZ (CARS-43-KXJ12) and LZ (CARS-43-SYZ14) and National Natural Science Foundation of China (32673780) to HZ. The experimental colonies were kindly provided by Xizang Ye Ban Feng Sheng Agriculture Technology Development Co., Ltd URL: http://www.ybfshoney.com. All funders had no role in the study, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The honey bee (Apis mellifera) is the most important managed global pollinator of both agricultural crops [1] and natural habitats [2], in part due to the worldwide decline of terrestrial insects [3]. Honey bees’ contribution to the annual global value of pollinations is estimated at US$235 to US$577 billion [4]. Over the past 60 years, a leading cause of honey bee death is the ectoparasitic mite Varroa destructor. It is the lethal combination of rapid mite population growth and its ability to transmit viral pathogens [5], that can cause colony death within three years in temperate regions, and sooner in warmer regions. Therefore, beekeepers use chemical miticides to control mite populations; despite these treatments, high colony losses persist, with V. destructor identified as the major cause [6].
In China, V. destructor was first recorded in A. mellifera colonies in 1955 and caused large losses as it spread across the country [7]. Miticide use became widespread with beekeepers currently treating colonies up to five times a year depending on location [8]. Thirty A. mellifera colonies were moved up onto the Tibetan Plateau to develop beekeeping in this region in 2004, 20 years later, the number of colonies has increased to over 12,000. Unexpectedly, migratory beekeepers reported that V. destructor appeared to disappear within a year after infested colonies were transported to the Tibetan Plateau from lower altitudes. Colonies in Tibet have not been treated since 2006, and no mite-related problems are reported. This is the first time V. destructor has ever been known to disappear from colonies. Thus, the aim of this study was to confirm the phenomenon and uncover the mechanism which could lead to new and innovative ways of mite control.
Results
Effect of altitude on V. destructor population growth
In May 2024, 60 colonies from the same source apiary in Xishuangbanna, Southern China (S1 Fig) were divided into four groups of 15 colonies, and their mite infestation of the worker brood determined (S2 Data). The group with the highest infestation was moved to Lhasa, Tibet (3,637 m) to increase the chances of finding infested cells, while the group with the lowest mite infestation remained at Xishuangbanna (554 m), to prolong the time to colony collapse. The other two groups were moved to Kunming (1,966 m) and Nyingchi (2,928 m) (S1 Fig). However, the initial pupal infestation rate had no significant effect on subsequent mite reproduction in singly infested cells or on the altitude-associated reduction in mite reproduction (see below). Typical seasonal colony-size patterns at the four locations are shown for context in Fig 1D. Worker brood showed similar stage-specific development in Xishuangbanna (554 m) and Lhasa (3,637 m), with a mean egg-to-adult emergence time of approximately 21 days at both sites and across the 1-, 3- and 12-month assessments after transfer to Lhasa (n = 15 workers per group and time point from three colonies; S1 and S2 Data).
Fig 1. The effects of altitude on Varroa destructor reproduction.
(A) Average V. destructor mite infestation rates in the worker brood at 0, 2 and 4 months after translocation. The colony number is given above the standard deviation (SD) bars. (B) Development of V. destructor offspring at the four different altitudes. Red crosses denote dead offspring that failed to develop further. (C) The average potential number of mature female daughters at bee emergence per reproductive cycle (R) in cells that contained a mature male mite in each group. (D) Illustrative estimates of typical seasonal colony size under normal management at each site. (E) Hypoxia experiment conducted at an altitude of 88 m showing the experimental setup, and the effect on the mite’s reproductive ability and offspring development under normoxia conditions (ambient partial pressure of oxygen [pO2] inside hives: ~20.4 kilopascals [kPa]) and hypoxia (~13.5 kPa pO2). Honey bee developmental stages are abbreviated as follows: L, stretched larva; WP, white-eyed pupa; PRP, pink-eyed pupa; PP, purple-eyed pupa; YX, yellow-thorax pupa; GX, gray-thorax pupa; and NE, resting pupa. The data underlying this figure can be found in S1 and S2 Data.
The infestation rate of the worker brood was recorded at two and four months and again one year after the colony translocation (S1 and S2 Data). Despite an initial brood infestation rate of 9.3% in the Lhasa group, infestation declined significantly to 3.3% after two months, 1.6% after four months and 0% after one year (Holm-adjusted p = 0.029, 0.007 and 0.001, respectively; Fig 1A). In contrast, infestation increased significantly in Xishuangbanna (from 1.0% to 14.1%) and Kunming (from 2.8% to 21.8%) after two months (both adjusted p < 0.001), and in Nyingchi after four months (from 3.4% to 10.6%; adjusted p = 0.002). A pilot study in 2023 also found that when 15 colonies were moved from Baoji (1,023 m) to Shannan (3,587 m), the infestation rate declined significantly from 3.6% one month after arrival to 1.2% after three months (adjusted p < 0.001; S1 Data).
After four months, only one of the 29 colonies at the two higher altitude sites (Lhasa and Nyingchi) had died, compared with 25 of the 30 colonies at the lower altitudes (Xishuangbanna and Kunming), most likely due to high mite levels (Fisher’s exact test, p < 0.001; Fig 1A). By one year, all colonies in Nyingchi had died, whereas 14 of the 15 colonies transferred to Lhasa remained alive and mite-free. After 2 years, 13 of the 15 Lhasa colonies remained alive without any miticide treatment or supportive management other than scheduled monitoring.
To assess whether external climatic variation across sites could confound these results, we measured brood-nest temperature and relative humidity in three colonies at two representative sites, Lhasa and Xishuangbanna (S2 Fig). Although the annual climatic conditions at each site are very different (S3 Fig), honey bee colonies maintain a highly stable brood-nest microclimate in which V. destructor reproduces, typically between 33 and 36 °C and ~65%–75% relative humidity, regardless of external fluctuations [9,10]. Consistent with this, mean brood-nest temperatures were similar between the highest and lowest study sites, ranging from 34.1 to 34.2 °C in Lhasa and 34.4 to 34.6 °C in Xishuangbanna (S2 Fig). Relative humidity also showed comparable values, ranging from 68.5% to 71.8% in Lhasa and 67.1% to 71.6% in Xishuangbanna.
Effect of altitude on V. destructor reproduction
At each location, we measured the reproductive ability of V. destructor based on a minimum of 218 infested cells per group distributed across the seven different honey bee developmental stages (Fig 1). We found altitude had a significant negative effect on the ability of V. destructor to produce (mated) adult female offspring per reproductive cycle (R) (LRT χ2 = 79.24, df = 3, p < 0.001; Fig 1C), whereas initial pupal infestation rate had no significant effect on the R value (LRT χ2 = 1.15, df = 1, p = 0.283).
The negative effect of altitude on V. destructor reproduction was due to a strong decline in mite fecundity (estimate = –0.3226 ± 0.0343, z = –9.394, p < 0.001; Fig 2B) and the delayed or arrested development of the offspring (Figs 1B and 2A). That is, in Lhasa and Shannan, mites produced a maximum of two eggs that rarely developed into protonymphs (Figs 2 and S4). Across all three translocation experiments, and at the two high-altitude sites (3,587 and 3,637 m), no potential mated adult female offspring were observed in any infested cell containing a yellow‑thorax honey bee pupa or older (n = 178; S1 Data).
Fig 2. Effect of altitude on Varroa destructor reproduction.
(A) The development of V. destructor offspring (left column) and their proportion within the cell (right column) at seven different developmental stages in honey bee worker pupae at different altitudes (3,637 m: Lhasa; 2,928 m: Nyingchi; 1,966 m: Kunming; 554 m: Xishuangbanna) from the translocation experiment, R values are as defined in Fig 1. (B) GLMM analysis of the effect of altitude on the number of V. destructor offspring produced per reproductive mite (data from the translocation experiment and the 2023 pilot studies are combined). (C) GLMM analysis of the number of V. destructor offspring produced per reproductive mite per cycle at three, six, nine, and 12 days after arriving in Lhasa. Treatment (days): LRT χ2 = 67.837, df = 4, p < 0.001; pupal infestation rate: LRT χ2 = 0.217, df = 1, p = 0.6413). The data underlying this figure can be found in S1 and S2 Data.
To investigate how rapidly this change in the mite’s reproductive ability occurred after arrival at high altitude, as well as the possible effect of colony transportation, we measured mite reproduction every 3 days after the colonies arrived in Lhasa. We found that mites at three days post‑transport already had very poor reproduction (R = 0.2) with more offspring arrested at the egg stage (S5 Fig), and a significant reduction in fecundity was already evident (Tukey-adjusted p < 0.0001; Fig 2C). From day 6 onward, mother mites only laid one or two eggs that failed to develop further (R = 0), as had previously been seen after longer periods (12 days to four months) (Figs 2C and S5). This pattern was consistent with our 2023 Lanzhou to Lhasa pilot study (S6 Fig). Hence, the effect of altitude on the mite’s reproductive ability was very rapid.
Low altitude hypoxic experiment
To confirm that the reduction in oxygen (hypoxia) at Lhasa (pO2: ~13.6 kPa) was the primary cause of reproductive failure in V. destructor, in 2024 we artificially reduced the ambient pO2 inside hives to 13.5 ± SD 2.3 kPa in five modified hives kept at an altitude of 88 m (S2 Data and S7 Fig). This was done by regulating a mixture of nitrogen and oxygen gases within the modified hives in which the bees were allowed to forage freely and observed the mite reproduction after 11 days. Again, we found that V. destructor only produced an average of 1.7 offspring, that was significantly fewer than the 3.9 produced by the controls (Tukey-adjusted p < 0.0001 versus control). Almost all eggs in the hypoxic colonies failed to develop further as seen previously in Lhasa (Figs 1E and 3), whereas in the control colonies normal mite reproduction was again seen (Figs 1E and 3). A pilot study in 2023 using six colonies similarly exposed to hypoxia (pO2: 14.3 ± SD 3.1 kPa; S2 Data) for six days also found that mite fecundity dropped to 2.1 eggs per reproductive mite (Tukey-adjusted p < 0.001 versus control), and offspring development was again severely arrested (Figs 3B and S8). In all honey bee colonies used, workers assessed at 11 days post‑capping were at the newly emerged adult stage in both the control and hypoxia treatment (n = 15, from five colonies per treatment), indicating no obvious delay in worker brood development under hypoxia at this assessment time point.
Fig 3. The effect of oxygen concentration on Varroa destructor reproduction.
(A) The number (left column) and proportion (right column) of mite offspring under control (~20.4 kPa pO2) and hypoxic conditions (~13.5 kPa pO2) for 11 days. R values are as defined in Fig 1. Mite offspring/honey bee developmental stages are as defined in Fig 2. (B) A GLMM analysis comparing the effect of hypoxic conditions on the number of V. destructor offspring per productive mother mite under control and hypoxic conditions after six and 11 days (data combined from main experiment and the 2023 pilot experiment). Treatment: LRT χ2 = 25.2057, df = 2, p < 0.001; pupal infestation rate: LRT χ2 = 0.0026, df = 1, p = 0.9594). The data underlying this figure can be found in S1 and S2 Data.
Discussion
This is the first time a hypoxia-related physiological limit that affects V. destructor reproduction has been discovered. The resulting disappearance of the mites from honey bee colonies opens new potential avenues for novel control methods.
Honey bees are known to forage up to altitudes of 4,100 m [11], which is consistent with the estimated critical pO2 for honey bee flight (12.15 kPa), allowing flight to an elevation of approximately 4,500 m [12]. As honey bee worker offspring development was not substantially delayed by altitude or hypoxia, the negative effect on V. destructor is unlikely to be mediated through impaired bee development. The similarity of brood nest temperature and humidity across sites (S2 Fig) further argues against these factors as major drivers of the observed altitudinal pattern. Although transportation may transiently stress honey bee colonies, long-distance movement has been routinely used in migratory beekeeping for decades and has not been consistently associated with reduced V. destructor prevalence [13]. Furthermore, the lower-altitude groups in our study were also transported by truck, and reproductive failure persisted for several months after the Lhasa colonies became stationary (Figs 2C and S6); and the same phenotype was reproduced in the low-altitude hypoxia experiment without colony transportation. Transportation is therefore unlikely to explain the observed reproductive failure. Instead, the systematic decline in oxygen partial pressure with increasing altitude provides a more parsimonious explanation for the reproductive failure of V. destructor at high elevations.
The sensitivity of V. destructor reproduction to reduced pO2 observed here may be related to the mite’s respiratory system, since each of its two tracheal networks opens to the exterior through a single spiracle on the ventrolateral surface [14]. Adult females also possess mobile peritremes that facilitate respiration while submerged in liquid brood food [15]. Although muscles directly controlling the spiracles or peritremes have not been detected, adult females have been reported to actively ventilate their tracheae by changing their body volume [16]. This movement of air represents a form of convective ventilation similar to that described in small insects [17]. These respiratory adaptations may help explain the continued presence of adult foundresses in colonies for several months under hypoxic conditions, as observed in our translocation experiments (Fig 1A and S1 Data). However, the ability of adult foundresses to survive does not necessarily mean that reproduction and offspring development have the same tolerance to reduced oxygen. This distinction is also seen in Tetranychus urticae [18] and Drosophila melanogaster [19], in which fecundity is more sensitive to reduced oxygen than adult survival. High elevation has similarly been associated with reduced fecundity and slower larval growth in the willow leaf beetle Chrysomela aeneicollis [20]. Eggs also appear particularly sensitive to hypoxia, experiencing delayed development and reduced survival [21]. Notably, newly hatched larvae of V. destructor lack tracheae and rely on cutaneous respiration, with the tracheal system forming only after the first moult [22]. This developmental feature may contribute to offspring vulnerability under oxygen-limited conditions, consistent with our observation that most offspring in Lhasa (~3,650 m) failed to develop beyond the prolarval stage. Whether these effects reflect maternal energy limitation, constraints on offspring development, mitochondrial function, oxygen sensing, metabolic regulation, or other physiological pathways [23,24] remains to be determined.
The discovery of this oxygen‑sensitive reproductive limit has immediate practical implications. As a proof of concept, we have tested a simple, low‑cost and non-chemical approach: by reducing the hive entrance opening at night, we allowed the bees’ own respiration to lower the internal oxygen concentration, which partially suppressed V. destructor reproduction in a pilot experiment. Ultimately, new non-chemical methods of controlling the mite will benefit both the beekeeping industry and our own food security.
Materials and methods
Study species and sites
The study used Apis mellifera ligustica honey bees, the most common sub-species managed in China [7]. All study colonies were originally treated regularly with miticides to control the V. destructor mite but were stopped at least two months prior to the start of each experiment to allow the mite populations to increase naturally. Throughout each study, no miticides were used, and the colonies were allowed to persist or die. Each study colony consisted of a lower box containing one or two queens, brood and pollen and an upper box that contained the honey. The two boxes were separated by a queen excluder that allowed workers but not the queen to enter the top box. When two queens were present, they were separated by a solid divider, but the workers could mix freely in the upper box. This type of two-queen system is commonly used throughout China [25]. The study moved colonies to a range of locations (S1 Fig) that were situated at different altitudes (S2 Data). Barometric pressure at each study site was measured using DYM3 and DYM3-1 aneroid barometers, and the corresponding atmospheric oxygen partial pressure was calculated from the measured barometric pressure and the atmospheric oxygen fraction. For each location, the typical colony size was estimated over the year using the number of frames covered by bees each month, which were then multiplied by 3,000 to obtain the total number of adults in the colony [26] (Fig 1D). To assess whether external climatic variability could confound our interpretation of altitude-related effects, brood-nest temperature and relative humidity were recorded in three colonies at Lhasa (3,637 m) and three colonies at Xishuangbanna (554 m) using temperature and relative-humidity data loggers (RS-WS-4G-C3H, Renke, China). The probes were positioned against the brood comb, while the main logger units remained outside the hives. Measurements were recorded at five-minute intervals.
To compare worker brood development between Xishuangbanna and Lhasa, queens were confined for six hours to obtain an even-aged brood. At each assessment, five developing workers were examined daily in each of three replicate colonies until adult emergence. The assessment was conducted in Xishuangbanna, and repeated at 1, 3, and 12 months after arrival in Lhasa.
The reproductive cycle of the honey bee ectoparasitic mite V. destructor is well established [27], and a subsequent study [28] has indicated that no detectable change has occurred in the reproductive cycle over at least the past 24 years. Honey bees regulate the temperature of their brood nest between 33 and 36 °C [9,10], so developmental times are directly comparable across different studies. Briefly, the foundress V. destructor invades a worker cell containing a fifth-stage honey bee larvae one day before the cell is sealed with a wax cap. After cell caping, the foundress mite feeds on the bee larvae and 60 hours later lays her first pro-larvae (referred to as an egg) that develops into a protonymph after 30 hours (male) or 22 hours (female). The first egg always develops into a male and the subsequent three to four eggs all develop into females. The protonymph stages lasts 52 hours (male) or 32 hours (female) followed by a deutonymph stage lasting 72 (male) to 80 (female) hours before molting into an adult. Brother–sister mating occurs within the cell, but only the adult females can survive outside the protection of the sealed cell living and feeding on the adult bees (phoretic stage). Within A. mellifera honey bee worker brood, there is only sufficient time for the male and up to three females to reach adulthood (Fig 1B). Using these details, the developmental stage of the honey bee pupae, presence of mite exuvia, and presence or absence of an adult male allowed us to calculate the potential number of adult female offspring when the honey bee pupa emerges (R) [27]. There is a trade-off between accuracy and sample size, so unless otherwise stated, we used pupae at the yellow-thorax stage or older to calculate R in cells that contained only a single foundress.
Effect of changing altitude on V. destructor population growth
In late May 2024, prior to translocation, 60 bee colonies were selected from an apiary in Xishuangbanna and divided into four groups of 15 colonies (S1 Data). Each colony had at least 15,000 adult bees, and their V. destructor infestation rate was determined using a minimum of 60 worker brood per colony, (Xishuangbanna control group: 1%, Kunming group: 3%, Nyingchi group: 3%, Lhasa group: 9%). On May 30, 15 hives were transported by truck to a new site in Xishuangbanna, arriving in the evening; another 15 hives arrived in Kunming the next day. On June 13, 14 hives were transported to Nyingchi (arriving in the afternoon of June 15, one hive was lost) and 15 hives arrived in Lhasa on June 16. After 2, 4 months and 1 year, the infestation rate of the worker brood and V. destructor reproductive status were measured by opening a minimum of 60 cells in all the colonies that contained sufficient brood, that is 900 cells per group of 15 colonies. In the Kunming group, after three months, all the colonies died most likely due to high V. destructor levels and attacks from hornets (Vespa velutina).
In addition to the translocation experiment, we conducted a pilot study in May 2023 where 15 colonies from Baoji (1,023 m) were transported to Shannan (3,587 m). At one and three months after arrival, we inspected a minimum of 100 worker brood cells per hive recording both mite’s infestation rate and reproduction.
Effect of increasing altitude on V. destructor reproduction
For the four-site comparison, mite reproduction was assessed before translocation in the Xishuangbanna source colonies and at 14–16 days after arrival in Kunming, 22–23 days after arrival in Nyingchi, and 12 days after arrival in Lhasa. Mite reproduction in Lhasa was additionally reassessed two and four months after arrival. To evaluate the reproductive status of V. destructor, a frame containing sealed worker brood was removed from the hive; then randomly chosen worker cells were opened using forceps, and the bee larvae or pupae extracted from the cell. Each bee and the interior of its cell were inspected for mites using a flashlight. If the cell was infested, the developmental stage of the bee was recorded using color changes of the pupa [29]. Then the number and developmental stages of the mite progeny (eggs, protonymphs, deutonymphs [female], adult males, and adult females) were recorded. At least 30 infested cells were collected for each of the seven sealed brood developmental stages, e.g., stretched larvae (40–80 hours post-capping [hpc]), white-eyed (80–110 hpc), pink-eyed (140–160 hpc), purple-eyed (150–160 hpc), yellow-thorax (190–230 hpc), gray-thorax (250–270 hpc), and resting pupa (270–280 hpc) [29].
In a 2023 pilot study, 20 colonies from Lanzhou (1,968 m) with a worker brood infestation rate of 35% (S1 and S2 Data) were transported to Lhasa. This provided additional data from a different year and location.
Assessment of the speed of altitude effects on V. destructor reproduction
To evaluate how rapidly high altitude impairs V. destructor reproduction, we performed time‑course experiments after the mite populations arrived in Lhasa. In the 2023 pilot (Lanzhou to Lhasa), 2 days after arrival mite reproduction was measured followed by inspections every 3 days for a total of eight checks, i.e., over a 23-day period. In 2024, using the colonies that arrived in Lhasa on June 16 the reproductive status of V. destructor was again examined every three days until June 28, i.e., four times over a 12-day period.
Low altitude hypoxic experiment
This study was conducted in Shaoxing, Zhejiang Province at an altitude of 88 m. The experiment followed a before‑after, within-hive design: the same five modified hives were first assessed under ambient conditions to obtain a normoxic baseline, followed by nitrogen infusion to establish hypoxia in the same hives. Three-frame beehives were adapted by inserting tubes into the lower corner through which the gas mixture entered the hive at it’s upper corner, diagonally opposite the side of the hive for ventilation. The hive entrance remained open throughout the experiment, allowing air exchange and unrestricted movement and foraging of the bees. Gas was supplied continuously from a cylinder to the inlet tube through a pressure-reducing regulator (A-1H, Zhongke Aier, China). Inside the hive, two monitoring devices (model VMS-3001-MG111-4G, Saien, China) were placed in the opposite diagonal corners of each hive to measure temperature, relative humidity, oxygen (vol%) and carbon dioxide concentration. The devices incorporated an electrochemical O2 sensor, a non‑dispersive infrared (NDIR) CO2 sensor, and an integrated temperature/humidity sensor chip and were factory‑calibrated before deployment. Measurements were transmitted continuously to an online monitoring platform and logged at 10-min intervals. Gas flow was adjusted manually at the regulator whenever the in-hive O2 mole fraction moved outside the target range of 12%–15%, thereby maintaining the intended hypoxic conditions. Oxygen vol% readings were subsequently converted to pO2 (kPa) using the concurrent barometric pressure. Cling film was wrapped around the inner cover and the hive body (excluding the top cover) to minimize gas leakage. A single entrance allowed the honey bees to forage normally. Into each adapted beehive, we placed a queen, bees and three frames, one containing sealed worker pupae with a V. destructor infestation rate of 20%–30%, a frame containing eggs and young larvae, and one frame containing honey and pollen. After setting up, the hives were left undisturbed for 6 hours to allow the colony to settle. In 2023, the cylinder contained 100% nitrogen. Using the monitoring-and-adjustment procedure described above, the mean in-hive pO2 was14.3 ± SD 3.1 kPa (S1 and S2 Data) across all six boxes. However, after 6 days, one colony had deserted their colonies, so mite reproduction was measured after 6 days in five colonies. In August 2024, the source gas was changed to a pre-mixed mixture of 5% O2 and 95% N2. The mean in-hive pO2 was 13.5 ± SD 2.3 kPa across the five hives (S1 and S2 Data). The mite reproductive status was checked after 11 days in all five colonies as the bees did not abscond.
Data processing
To reduce variability, all analyses excluded any multiple‑invade cells (4.7%–11.3% of the population). For fecundity (the number of mite offspring produced per reproductive mite per cycle), non‑reproducing mites (3.3%–8.5% of the population) were also removed; when calculating R, non‑reproducing mites were included. To improve accuracy, only honey bee yellow‑thorax pupae and later pupal stages were used for calculating R, whereas only gray‑thorax pupae and later pupal stages were used to calculate fecundity. All analyses used Poisson GLMMs with colony number as a random intercept and pupal infestation rate as a covariate (glmmTMB v1.1.14, R v4.5.2). Post‑hoc Tukey comparisons used emmeans (v2.0.2). In every case, the focal predictor (altitude or treatment) × infestation rate interaction was non‑significant and removed by backward selection (LRT, AIC). R value: Altitude was a categorical fixed factor. As all R values at 3,587 and 3,637 m were zero (causing estimation instability), we added one to every count (R + one). Fecundity: Data from all translocation experiments was combined (six altitudes: 554, 1,966, 1,968, 2,928, 3,587, 3,637 m). Altitude was a continuous predictor (scaled to unit variance), and days after arrival in Lhasa was a categorical fixed factor for the speed of altitude effects on V. destructor reproduction experiment. Hypoxia experiment: Treatment was a categorical fixed factor.
Changes in worker-brood infestation rates were analyzed separately for each site using two-sided paired Wilcoxon signed-rank tests, with p-values adjusted across the eight comparisons using the Holm method. Only colonies with measurements available at both time points were included. Colony mortality after four months was compared between the two higher-altitude and two lower-altitude sites using a two-sided Fisher’s exact test.
Supporting information
S4 Fig. Effect of altitude on Varroa destructor reproduction in pilot study.
The development of mite offspring (left column) and their proportion within the cell (right column) at seven different developmental stages in honey bee worker pupae at different altitudes (3,637 m: Lhasa, 2023; 3,587 m: Shannan; 1,968 m: Lanzhou) from the pilot experiment (S1 and S2 Data). R values are as defined in Fig 1. Honey bee pupal stages are as defined in Fig 2.
https://doi.org/10.1371/journal.pbio.3004091.s004
(TIF)
S5 Fig. The speed of altitude effects on Varroa destructor reproduction.
The development of mite offspring (left column) and their proportion within the cell (right column) at seven different developmental stages in honey bee worker pupae at 3, 6, 9, and 12 days after arriving in Lhasa from the main experiment (S1 and S2 Data). R values are as defined in Fig 1. Honey bee pupal stages are as defined in Fig 2.
https://doi.org/10.1371/journal.pbio.3004091.s005
(TIF)
S6 Fig. The speed of altitude effects on Varroa destructor reproduction in pilot study.
(A) The development of mite offspring (left column) and their proportion within the cell (right column) at seven different developmental stages in honey bee worker pupae at 2, 5, 8 and more than 11 days after arriving in Lhasa from the pilot experiment (S1 and S2 Data). R values are as defined in Fig 1. Honey bee pupal stages are as defined in Fig 2. (B) GLMM analysis of the number of mite offspring produced per reproductive mite per cycle at two, five, eight and more than 11 days after arriving in Lhasa from the pilot experiment. Treatment (days): LRT χ² = 44.664, df = 4, p < 0.001; pupal infestation rate: LRT χ² = 0.234, df = 1, p = 0.6287).
https://doi.org/10.1371/journal.pbio.3004091.s006
(TIF)
S7 Fig. Oxygen partial pressure (pO2) in artificial hypoxic hive.
A typical trace of the oxygen partial pressure (pO2 kPa), calculated from the sensor readings using the measured barometric pressure recorded by one of the monitors in the artificially modified hives in 2024 over an 11-day period (S1 Data).
https://doi.org/10.1371/journal.pbio.3004091.s007
(TIF)
S8 Fig. The effect of oxygen concentration on Varroa destructor reproduction in pilot study.
The number (left column) and proportion (right column) of mite offspring under control (20.4 kPa pO2) and hypoxic conditions (14.3 kPa pO2 for six days) (S1 and S2 Data). R values are as defined in Fig 1. Honey bee pupal stages are as defined in Fig 2.
https://doi.org/10.1371/journal.pbio.3004091.s008
(TIF)
Acknowledgments
We thank Liangfu Zhang, Qiwei Gu for help with managing the study colonies. Also the teams led by Zhengwei Wang at Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, and Jun Guo at Kunming University of Science and Technology for translocation site support. We thank Séverine Kotrschal and Delphine Panziera at Wageningen University and Research, the Netherlands for statistical assistance, Nihong Zhou, Qiulan Zheng, Meifei Zhu, Yanling Xie, Bin Yuan from Zhejiang University, and Yuan Zhao, Puchi Bianba from Xizang Academy of Agriculture and Animal Husbandry Sciences for experimental support.
References
- 1. Calderone NW. Insect pollinated crops, insect pollinators and US agriculture: Trend analysis of aggregate data for the period 1992-2009. PLoS One. 2012;7(5):e37235. pmid:22629374
- 2. Hung K-LJ, Kingston JM, Albrecht M, Holway DA, Kohn JR. The worldwide importance of honey bees as pollinators in natural habitats. Proc Biol Sci. 2018;285(1870):20172140. pmid:29321298
- 3. van Klink R, Bowler DE, Gongalsky KB, Swengel AB, Gentile A, Chase JM. Meta-analysis reveals declines in terrestrial but increases in freshwater insect abundances. Science. 2020;368(6489):417–20. pmid:32327596
- 4. Potts SG, Imperatriz-Fonseca V, Ngo HT, Aizen MA, Biesmeijer JC, Breeze TD, et al. Safeguarding pollinators and their values to human well-being. Nature. 2016;540(7632):220–9. pmid:27894123
- 5. Wilfert L, Long G, Leggett HC, Schmid-Hempel P, Butlin R, Martin SJM, et al. Deformed wing virus is a recent global epidemic in honeybees driven by Varroa mites. Science. 2016;351(6273):594–7. pmid:26912700
- 6. Thompson J. Scientists identify culprit behind biggest ever U.S. honey bee die-off. Science. 2025;389(6756):114–5. pmid:40638723
- 7. Zheng H, Cao L, Huang S, Neumann P, Hu F. Current status of the beekeeping industry in China. Asian Beekeeping in the 21st Century. Springer Singapore; 2018. p. 129–58.
- 8. Qiu R, Shen J. On the control of varroa mites in Italian honey bees. Apiculture of China. 2021;72(02):24–5.
- 9. Simpson J. Nest climate regulation in honey bee colonies. Science. 1961;133(3461):1327–33.
- 10. Stabentheiner A, Kovac H, Brodschneider R. Honeybee colony thermoregulation: Regulatory mechanisms and contribution of individuals in dependence on age, location and thermal stress. PLOS ONE. 2010;5(1):e8967.
- 11. Thapa R, Aryal S, Jung C. Beekeeping and honey hunting in nepal: Current status and future perspectives. Asian Beekeeping in the 21st Century. Springer Singapore; 2018. p. 111–27.
- 12. Woods HA, Guzman PE, Cheviron ZA. The oxygen and pressure physiology of upper elevational range limits in insects. Functional Ecology. 2026;40(10):3020–33.
- 13. Martínez-López V, Ruiz C, De la Rúa P. Migratory beekeeping and its influence on the prevalence and dispersal of pathogens to managed and wild bees. Int J Parasitol Parasites Wildl. 2022;18:184–93. pmid:35663725
- 14. Richard D, Colin ME, Lhomme M. Anatomical organization of the tracheal system of Varroa jacobsoni (Acari: Varroidae). Exp Appl Acarol. 1990;9(1–2):63–72.
- 15. Pugh PJA, King PE, Fordy MR. The respiratory system of the female Varroa jacobsoni (Oudemans): Its adaptations to a range of environmental conditions. Exp Appl Acarol. 1992;15(2):123–39.
- 16. Akimov IA, Yastrebtsov AV. Respiratory system of the mite Varroa jacobsoni (Parasitiformes, Varroidea). Vestnik Zoologii. 1984;18(3):45–9.
- 17.
Harrison JF, Woods HA, Roberts SP. Ecological and Environmental Physiology of Insects. Oxford University Press; 2012.
- 18. Yamakawa S, Ohyama K, Yamori W, Suzuki T. Effects of anoxia and hypoxia on the two-spotted spider mite, Tetranychus urticae (Acari: Tetranychidae). Appl Entomol Zool. 2018;53(4):535–41.
- 19. Peck LS, Maddrell SHP. Limitation of size by hypoxia in the fruit fly Drosophila melanogaster. J Exp Zool A Comp Exp Biol. 2005;303(11):968–75. pmid:16217805
- 20. Dahlhoff EP, Dahlhoff VC, Grainger CA, Zavala NA, Otepola‐Bello D, Sargent BA, et al. Getting chased up the mountain: High elevation may limit performance and fitness characters in a montane insect. Funct Ecol. 2019;33(5):809–18.
- 21. Harrison JF, Greenlee KJ, Verberk W. Functional hypoxia in insects: Definition, assessment, and consequences for physiology, ecology, and evolution. Annu Rev Entomol. 2018;63:303–25. pmid:28992421
- 22.
Sadov AV. The respiratory apparatus of the female Varroa jacobsoni mite. Veterinariya. 1980;(11):43–7.
- 23. Hetz SK. The role of the spiracles in gas exchange during development of Samia cynthia (Lepidoptera, Saturniidae). Comp Biochem Physiol A Mol Integr Physiol. 2007;148(4):743–54. pmid:17855137
- 24. Khan S, Staples AE. Mechanisms of insect respiration. Nat Rev Phys. 2025;7(3):135–48.
- 25. Zheng H-Q, Jin S-H, Hu F-L, Pirk CWW, Dietemann V. Maintenance and application of multiple queen colonies in commercial beekeeping. J Apicult Res. 2009;48(4):290–5.
- 26. Jeffree EP. A photographic presentation of estimated numbers of honeybees (Apis melufera L.) on combs in 14 × 8½ inch frames. Bee World. 1951;32(12):89–91.
- 27. Martin SJ. Ontogenesis of the mite Varroa jacobsoni Oud. in worker brood of the honeybee Apis mellifera L. under natural conditions. Exp Appl Acarol. 1994;18(2):87–100.
- 28. Oddie M, Büchler R, Dahle B, Kovacic M, Le Conte Y, Locke B, et al. Rapid parallel evolution overcomes global honey bee parasite. Sci Rep. 2018;8(1):7704. pmid:29769608
- 29. Dietemann V, Nazzi F, Martin SJ, Anderson DL, Locke B, Delaplane KS, et al. Standard methods for varroa research. J Apicult Res. 2013;52(1):1–54.






















English (US) ·
French (CA) ·